Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
The Career Achievements of National Research Service Award Postdoctoral Trainees and Fellows: 1975–2004
May 2006
prepared for
Office of Extramural ProgramsOffice of the Director
National Institutes of Health
by
Richard Mantovani, Ph.D.Mary V. Look, Ph.D.
Emily Wuerker
7315 WISCONSIN AVENUE • SUITE 400W • BETHESDA, MD 20814
ACKNOWLEDGMENTS
This study represents an effort by the National Institutes of Health (NIH) to understand outcomes associated with Kirschstein-NRSA postdoctoral support. The study was originally guided by NIH’s Committee on Research Training Assessment. Dr. Georgine Pion, at that time of Vanderbilt University, was responsible for developing a research design and for overseeing the construction of the data files by ORC Macro. In 2004 ORC Macro received a contract from the NIH Office of the Director Office of Extramural Research to complete the study using the original population and datasets. In the current study, Dr. Walter Schaffer took a leadership role in reviewing the report as it progressed and provided guidance to the authors throughout the study. Dr. Walter Goldschmidt and Dr. William McGarvey have contributed throughout the process by reviewing the methodology and results and providing critical feedback. We thank all three of them for providing important input on the program, our interpretations of the data, and the methodology employed, as well as for their patience.
We would like to thank Dr. Pion for providing information on the original data collection and research design, helping us meet some of the methodological and data challenges that we encountered, and reviewing drafts of the report. We would also like to thank Dr. Charles Sherman for reviewing the report.
We would like to thank several ORC Macro staff for their contributions to the study: Mr. Don McMaster, for his coordination of data-gathering efforts and for reviewing the study, and Ms. Vaishali Joshi, Ms. Shirley Zhou, and Mr. Philip Schaeffer for preparing the data for the analyses and tables. Finally, we thank Ms. Susan Wimmer and Ms. Cindy Young-Turner for providing excellent editorial assistance throughout the many drafts of the study report.
ORC Macro i
CONTENTS
EXECUTIVE SUMMARY.............................................................................................................1
1. AN OVERVIEW OF THE KIRSCHSTEIN-NRSA PROGRAM.............................................7
1.1. Changes in Kirschstein-NRSA Postdoctoral Policies and Practices................................71.1.1. Maximum Duration of Support.............................................................................81.1.2. Financial Support Package....................................................................................91.1.3. Service Payback Requirement............................................................................10
1.2. Growth of Postdoctoral Training Programs....................................................................111.3. Kirschstein-NRSA Postdoctoral Training Mechanisms.................................................12
1.3.1. T32 Traineeship Program...................................................................................131.3.2. F32 and F35 Fellowship Programs.....................................................................13
1.4. Major Study Questions and Methodology......................................................................151.4.1. Study Populations...............................................................................................18
1.4.1.1. Analysis of Professional Doctorates......................................................191.4.2. Selection of Study Populations...........................................................................19
1.4.2.1. Selection of the Kirschstein-NRSA Study Population..........................201.4.2.2. Kirschstein-NRSA and Other Study Group Subpopulations—SDR
Analyses................................................................................................201.4.3. Data Files and Data Elements.............................................................................21
1.5. Focus of the Study..........................................................................................................21
2. PREDOCTORAL EXPERIENCES AND BACKGROUND CHARACTERISTICS OF NEW RESEARCH DOCTORATES.......................................................................................23
2.1. Considerations in Characterizing Research Doctorate Backgrounds.............................232.2. Data Sources and Definitions.........................................................................................242.3. Do Kirschstein-NRSA Postdoctoral Support Recipients Receive Their Research
Doctorates From More Research-Intensive Institutions?...............................................262.3.1. Level of Research Intensity as Measured by the Carnegie Classification..........272.3.2. Level of Research Intensity as Measured by NIH Research Funding................30
2.4. To What Extent Did Individuals Receive Kirschstein-NRSA Support During Their Graduate Training?.........................................................................................................32
2.5. How Much Time Did Individuals Spend in Pursuit of the Doctorate?...........................352.5.1. Time to the Doctorate.........................................................................................362.5.2. Enrolled Time to the Doctorate..........................................................................38
2.6. What Are the Demographics of Doctorate Recipients?..................................................412.7. Did New Doctorate Recipients Have Plans To Continue With Postdoctoral Study?.....462.8. Summary and Conclusions.............................................................................................48
2.8.1. Summary.............................................................................................................482.8.2. Conclusions.........................................................................................................53
3. THE NATURE OF KIRSCHSTEIN-NRSA SUPPORT.........................................................54
3.1. What Kind of Kirschstein-NRSA Support Was Received for the First and Subsequent Postdoctoral Positions?...................................................................................................55
ORC Macro ii
3.1.1. Subsequent Kirschstein-NRSA Support Patterns for Individuals Who Initially Applied for and Received a Fellowship..............................................................62
3.1.2. Subsequent Kirschstein-NRSA Support Patterns for Individuals Who Initially Received Traineeships........................................................................................64
3.1.3. Subsequent Kirschstein-NRSA Support Patterns for Individuals Who Failed To Obtain Fellowships on Their First Attempt...................................................66
3.2. What Was the Length of Kirschstein-NRSA Postdoctoral Support?.............................683.3. In What Fields Were Recipients Doing Research While Receiving Kirschstein-NRSA
Support?..........................................................................................................................713.4. What Was the Research Intensity of the Institution at Which the Kirschstein-NRSA
Fellow or Trainee Conducted Research?........................................................................733.5. Summary and Conclusions.............................................................................................75
4. CAREER OUTCOMES OF KIRSCHSTEIN-NRSA POSTDOCTORAL RECIPIENTS.....77
4.1. Introduction.....................................................................................................................774.2. NIH Research Funding...................................................................................................79
4.2.1. Grant-Related Outcome Measures......................................................................804.2.2. Data Sources and Methodology..........................................................................814.2.3. The Effect of the Kirschstein-NRSA Program for Research Postdoctorates......82
4.2.3.1. Application and Award Rates................................................................834.2.3.2. Number of Awards................................................................................884.2.3.3. Award Rate............................................................................................904.2.3.4. Years to First Competing R01 or R01-Equivalent Award....................924.2.3.5. Summary...............................................................................................94
4.2.4. The Effects of the Kirschstein-NRSA Fellowship and Traineeship Programs on Research and Professional Doctorates...........................................................944.2.4.1. NIH Grant Application and Award Rates.............................................964.2.4.2. Number of Awards..............................................................................1014.2.4.3. Award Rates........................................................................................1034.2.4.4. Years to First Award...........................................................................1064.2.4.5. Overview.............................................................................................109
4.3. Publications...................................................................................................................1094.3.1. Outcome Measures, Methodology, and Data Sources......................................1104.3.2. Results...............................................................................................................110
4.3.2.1. Percentages of Research Doctorates Who Have Published.................1114.3.2.2. Mean Number of Publications.............................................................1144.3.2.3. Elapsed Years From Doctorate to First Publication............................1164.3.2.4. Citations...............................................................................................118
4.3.3. Regression Results on the Number of Publications..........................................1204.3.4. Summary...........................................................................................................121
4.4. Employment..................................................................................................................1214.4.1. Outcome Measures, Methodology, and Data Sources......................................1224.4.2. Results...............................................................................................................123
4.4.2.1. Employment Setting............................................................................1244.4.3. Academic Rank and Tenure Status...................................................................1264.4.4. Summary...........................................................................................................130
ORC Macro iii
4.5. Summary and Conclusions...........................................................................................1314.5.1. Conclusions.......................................................................................................137
CHAPTER 4 ADDENDUM: ANALYSIS OF NIH GRANT APPLICATION AND AWARD RATES FROM 1990 TO 2002..............................................................................................138
Conclusions............................................................................................................................143
5. SUMMARY AND CONCLUSIONS....................................................................................145
5.1. Study Outcomes............................................................................................................1455.1.1. Success in Obtaining NIH Research Grant Support.........................................1455.1.2. Success in Publishing in Peer-Reviewed Journals............................................1465.1.3. Success in Obtaining and Remaining in Research-Oriented Employment.......146
5.2. Study Groups................................................................................................................1475.2.1. Measuring the Effects of Outcomes..................................................................149
5.3. Results of the Outcomes Analysis................................................................................1505.3.1. Success of Research Doctorates in Obtaining NIH Research Grant Support...150
5.3.1.1. Overview.............................................................................................1515.3.2. Success of All Kirschstein-NRSA Postdoctorates in Obtaining NIH Research
Grant Support....................................................................................................1515.3.3. Publication Success of Research Postdoctorates..............................................152
5.3.3.1. Overview.............................................................................................1535.3.4. Employment Success of Research Postdoctorates............................................153
5.3.4.1. Overview.............................................................................................1535.4. Conclusions...................................................................................................................154
APPENDIXES
A: Predoctoral Experiences and Background Characteristics of New Research Doctorates TablesB: Nature of Kirschstein-NRSA Support TablesC: Career Outcomes TablesD: Regression Tables
ORC Macro iv
EXECUTIVE SUMMARY
OVERVIEW
Since its implementation in fiscal year (FY) 1975, the Ruth L. Kirschstein National Research Service Award (NRSA) program (Kirschstein-NRSA program) has focused on supporting predoctoral and postdoctoral training in a wide range of health-related disciplines. Two kinds of awards have been made: 1) institutional training grants to public and nonprofit private institutions for multidisciplinary training programs or targeted training programs in specific research areas and 2) fellowships to individuals for supervised study with a senior scientist at an institution of the fellow’s choice. Both awards are conferred after competitive peer review. Note that the first type of award is to the institution, which in turn appoints researchers to traineeship positions, while the second is to an individual who has competed with others for an award. Now that the Kirschstein-NRSA program has entered its fourth decade, there is a substantial history of individuals who received fellowship or traineeship support and then entered and prospered in research careers in academia or business. This study examines the effect of Kirschstein-NRSA postdoctoral support on research career outcomes and explores the evidence that such support produces individuals who have productive, research-focused careers. It is a followup to an earlier study by Garrison and Brown1 and corresponds to a more recent study by Pion2 on the career achievements of Kirschstein-NRSA postdoctorates.
METHODOLOGY
The study used information contained in the National Institutes of Health (NIH) Consolidated Grant Applicant File (CGAF) and Trainee and Fellow File (TFF) as well as the Doctorate Records File (DRF) and the longitudinal Survey of Doctorate Recipients (SDR). The timeframe for the study included individuals who received support from FY 1975 through FY 1992. FY 1992 was selected as the last year in order to provide at least 10 years after the completion of the postdoctorate in which to observe outcomes.
Three types of outcomes related to potential scientific contributions were explored: success in obtaining NIH research grant support, success in publishing in peer-reviewed journals, and success in obtaining and remaining in research-oriented employment. Focusing on these outcomes provides a perspective on how well individuals supported by Kirschstein-NRSA have performed throughout their careers compared with other equivalent groups of doctorates. We constructed six study groups to use in our analysis:
Kirschstein-NRSA fellows—Research doctorates who received a Kirschstein-NRSA fellowship through an F32 or F35 grant mechanism
1 Howard H. Garrison and Prudence W. Brown. (1986). The Career Achievements of NIH Postdoctoral Trainees & Fellows. NIH Program Evaluation Report, Committee on National Needs for Biomedical and Behavioral Research Personnel, Institute of Medicine. Washington, DC: National Academies Press.2 Georgine Pion. (2001). The Early Career Progress of NRSA Predoctoral Fellows and Trainees, http://grants.nih.gov/training/career_progress/.
ORC Macro 1
Kirschstein-NRSA trainees—Research doctorates who were appointed to a Kirschstein-NRSA fellowship by an institution with a T32 (training) grant
Kirschstein-NRSA trainee/fellows—Research doctorates who received both Kirschstein-NRSA traineeship and fellowship support
Unsuccessful fellowship applicants—Research doctorates who applied for but did not receive a Kirschstein-NRSA fellowship
Other non-Kirschstein-NRSA postdoctorates—Research doctorates who were appointed to a postdoctoral position supported by non-Kirschstein-NRSA sources such as the National Science Foundation
Other research doctorates—Research doctorates who had no indicated postdoctoral activity, including the submission of an application for a Kirschstein-NRSA fellowship
Another comparison used is between individuals with research doctorates and those with medical degrees, or those with dual (M.D./Ph.D.) degrees. This comparison was used to examine whether those with medical degrees have the same career profiles as those with research doctorates.
One critical issue in measuring differences in outcomes using the various study groups is the extent to which individual traits or predoctoral training experiences affect outcomes independent of Kirschstein-NRSA postdoctoral support: If the most highly qualified applicants received Kirschstein-NRSA support, any difference in outcomes may not be a result of postdoctoral training but may instead reflect selection factors. One approach to addressing the selection issue is to control for various factors, such as research intensity of the institution granting the doctorate; elapsed and enrolled time in pursuit of the doctorate; receipt of Kirschstein-NRSA predoctoral support; doctoral field; gender, marital status, and number of dependents; and postdoctoral plans.
In general, we found that there were differences between Kirschstein-NRSA postdoctorates and other research doctorates (including non-Kirschstein-NRSA postdoctorates) on almost all of these measures. Elapsed and enrolled time in pursuit of the doctorate was most useful in distinguishing between the study groups, and our analysis found that Kirschstein-NRSA fellows completed their doctoral requirements more quickly than trainees as well as other study groups. This may indicate that Kirschstein-NRSA fellows are able to advance their careers more quickly than individuals in the other study groups. A comparison of Kirschstein-NRSA fellows and trainees by months of support seemed to indicate that certain groups of fellows with more months of support were different from fellows with fewer months of support.
OUTCOMES ANALYSIS
SUCCESS IN OBTAINING NIH RESEARCH GRANT SUPPORT
We first examined success in obtaining R01 and R01-equivalent funding, which includes R01, R23, and R29 awards. The focus was on several grant-related outcome measures: percentage applying, percentage receiving awards, number of grants received, percentage of total applications that resulted in awards, and mean years until first award. The first analysis focused
ORC Macro 2
on research doctorates only, in order to allow for comparison between Kirschstein-NRSA and non-Kirschstein-NRSA postdoctoral populations. Our findings included:
Kirschstein-NRSA fellows performed better than most groups on most measures related to grant outcomes.
In contrast, Kirschstein-NRSA trainees performed at or below the level of unsuccessful fellowship applicants and other non-Kirschstein-NRSA postdoctorates.
Kirschstein-NRSA trainee/fellows performed close to fellows on most measures. One measure, years to first R01 award, seems to indicate that Kirschstein-NRSA postdoctoral
recipients experience a lag in obtaining R01 support when compared with other research doctorates who presumably become employed shortly after the completion of the doctorate.
This analysis indicates that although there is a difference between the Kirschstein-NRSA-supported study groups and the non-Kirschstein-NRSA study groups, this difference largely reflects the differences in the performance of Kirschstein-NRSA fellows and trainee/fellows. In other words, Kirschstein-NRSA trainees tend not to do as well as the other Kirschstein-NRSA groups in securing NIH grant funding and were more like the other comparison groups. There is also evidence that the Kirschstein-NRSA-supported groups secure R01 and R01-equivalent funding later than the other study groups.
The second analysis focused on all Kirschstein-NRSA postdoctorates, including professional doctorates. With regard to comparisons between Ph.D.’s, M.D.’s, and those with M.D./Ph.D.’s, we found that there were differences in success rates between the three groups:
M.D.’s seemed to have less success in securing R01 funding than Ph.D.’s and M.D./Ph.D.’s, particularly for Kirschstein-NRSA fellows.
For Kirschstein-NRSA fellows, Ph.D.’s and M.D./Ph.D.’s had higher award rates and fewer years to first award than those with only an M.D.
For Kirschstein-NRSA trainees, M.D./Ph.D.’s differentiated themselves from the other two degree groups, particularly in terms of securing R01 grant funding.
Overall, there is a persistent difference across Kirschstein-NRSA support groups on these measures, suggesting that although the type of degree makes a difference, the type of Kirschstein-NRSA support is critical to examining performance in terms of securing NIH grant funding.
SUCCESS IN PUBLISHING IN PEER-REVIEWED JOURNALS
For measures of publication success, we examined two cohorts of research doctorates: those who received their research doctorates in 1981–82 and 1987–88. Publication activity was examined from the year individuals received their research doctorates until 1995. The analysis focused primarily on the 1981–82 cohort, since they had a longer period in which to publish. Publication outcome measures included the number of publications, years to first publication, and average number of citations per publication.
ORC Macro 3
Kirschstein-NRSA postdoctoral support recipients were generally more successful on all publication outcome measures than other postdoctorates and other research doctorates.
There is little difference among Kirschstein-NRSA postdoctoral subgroups; however, there seems to be some evidence that Kirschstein-NRSA fellows experience a lag in publishing when compared with trainees.
Higher productivity was observed for Kirschstein-NRSA postdoctoral recipients who had 25–36 months of support; lower productivity was observed for those with 0–9 months of support.
SUCCESS IN OBTAINING AND REMAINING IN RESEARCH-ORIENTED EMPLOYMENT
Employment outcomes reflect a continuing interest and effort in research. We assessed various employment-related measures collected through the SDR 4, 8, and 12 years after individuals received their doctorates. Employment outcomes examined included employment in academic or business/industry settings and academic rank and tenure.
Kirschstein-NRSA postdoctorates were more likely to find employment in academic settings than the other study groups. They were also more likely to be employed in academic settings than business/industry settings. After 12 years, however, they were more likely to be located in business/industry settings than the other study groups.
After 4 years, Kirschstein-NRSA postdoctorates were not as advanced in their careers as other research doctorates or other non-Kirschstein-NRSA postdoctorates.
The differences between the non-Kirschstein-NRSA and Kirschstein-NRSA study groups grow smaller in subsequent time periods, thus indicating that Kirschstein-NRSA postdoctorates have accelerated career paths.
After 12 years, the differences between Kirschstein-NRSA postdoctorates and other non-Kirschstein-NRSA postdoctorates have decreased.
Although Kirschstein-NRSA postdoctorates tend to focus on academic employment, their career in this setting is delayed. In particular, other research doctorates who obtain academic positions achieve tenure and higher academic rank in a shorter time after receiving their doctorates than Kirschstein-NRSA postdoctorates. However, it should be noted that Kirschstein-NRSA postdoctorates, particularly fellows, seem to catch up with the other study groups. Kirschstein-NRSA fellows seem to achieve higher academic rank more quickly than trainees. Finally, Kirschstein-NRSA fellows and trainees are more likely to be located in business and industry settings after 12 years, while other non-Kirschstein-NRSA postdoctorates tend to find themselves in other organizational settings.
CONCLUSIONS
The major conclusion drawn from this study is that Kirschstein-NRSA fellows seem to differentiate themselves from those in the other study groups, including Kirschstein-NRSA trainees. Such success, however, is delayed relative to those who begin their research careers directly after receiving their doctorates. The difference between Kirschstein-NRSA fellows and
ORC Macro 4
trainees is important to note, since the outcomes of the latter group do not substantially differentiate themselves from other postdoctoral groups, such as those who applied unsuccessfully for a Kirschstein-NRSA fellowship. Because individuals take different postdoctoral support paths with differing lengths of support, the Kirschstein-NRSA program cannot be viewed as a unitary program. Capturing more detailed information about these paths would present a better sense of the kinds of research experiences provided and how they affect outcomes. For example, the study does not reflect on those individuals who receive support through R01 grants and various types of other NIH research grants as research associates. Thus, principal investigators receiving such support may select doctorates to work on their projects as employees. NIH currently does not collect easily accessible records on these individuals, but it has been speculated that this is a major source of support for those involved in postdoctoral studies and research.
One finding was an apparent lag among those receiving support from Kirschstein-NRSA training mechanisms. The lag effect suggests that Kirschstein-NRSA fellows and trainees start their academic careers later than others and also achieve tenure and higher academic rank later. This finding is distorted, however, particularly for those in the biomedical sciences, in which postdoctoral work has become the norm and relatively few individuals with research interests actually proceed into a research-oriented academic position after receiving the doctorate. Thus, it may be the case that individuals in comparison groups may receive academic appointments in institutions that have lower research expectations of faculty, and in these institutions they may achieve tenure more quickly. The issue here, as in much of the study, is identifying appropriate comparison groups.
Finally, a question could be raised about whether the findings in this study apply to individuals receiving their doctorates since 1992. In general, we believe that the findings apply to this more recent group of doctorates. However, it should be noted that very recent doctorates have less time to achieve the various outcomes studied here. On average, it takes a minimum of 5 years to receive NIH research funding, which means that a majority of those who received their doctorates since 1997 would not have even applied for NIH research funding. Results concerning intermediate or long-term outcomes become tenuous when we consider these more recent cohorts.
RECOMMENDATIONS
This analysis was limited in two regards: 1) the methodology and the comparison groups that underlie the methodology were formulated about a decade ago, leaving some current questions of interest potentially unaddressed, and 2) the data sources were limited largely to those assembled at that time. This study, however, pointed to certain methodological perspectives that may be useful in formulating future studies:
It is important to distinguish between Kirschstein-NRSA trainees and fellows. Outcomes for these two groups of individuals were very different and may suggest a strategy of identifying comparison groups for each group. One source of comparisons could include those individuals who receive support from R01 or other RPG grants. However, this would require identifying these individuals, which currently is not possible in the CGAF or TFF files or
ORC Macro 5
generally through the Information for Management, Planning, Analysis, and Coordination (IMPAC) II system.
Length of time receiving Kirschstein-NRSA support is important, especially for fellows. It seems that there are differences between Kirschstein-NRSA fellows who have fewer years of support and those with more years of support. Examining such information may suggest the optimal period for supporting fellows.
Emphasis should be placed on employment histories and longitudinal analysis. This study was cross-sectional, which meant that it was difficult to assess the complex career paths of individuals. For example, we found that each of the study groups included individuals who had very different experiences in applying for and obtaining support from NIH. In this context, support from R01 grants immediately after receipt of the doctorate should be taken into account as both a substitute for postdoctoral funding and a continuation to such funding.
It may be useful to identify and retrospectively analyze high achievers, both those with and without Kirschstein-NRSA support. A retrospective study may provide valuable insight on the program’s success in producing high achievers. Identifying these individuals and relating the program to their career paths may provide new insight on how the program can lead to significant career achievements.
A second set of recommendations relates to data that would enhance the analysis. In general, it would be useful to obtain data that are currently not accessible or easily available through current sources in order to promote analyses. These include:
Data that would allow for the tracking of foreign nationals. Foreign nationals are a growing segment of those who obtain Kirschstein-NRSA funding, but there are no data available on these individuals in the current DRF and SDR data files.
Data that would allow for the tracking of M.D.’s and other individuals with professional doctorates. These data have traditionally been gathered from sources such as the Association of American Medical Colleges or American Medical Association. For this study, obtaining information from these sources was problematic, which led to limited focus on these groups.
Tracking individuals working on research grants. This would be a necessary element for formulating comparison groups and formulating policy on training programs.
Development of better publication and employment outcomes. Publications data were particularly problematic because the current systems do not provide exact links to researchers. Individuals with similar names may be associated with the same article and thus increase measurement error. Also, authorship in publications may not be easily parsed in terms of the author’s contribution to that publication. In terms of employment, it may be important to identify the exact kinds of work in which individuals become involved, rather than a general question about whether they still engage in research. For example, the growth of private ventures in genetic engineering may pose more interest for some individuals than that of academia.
These recommendations would lay the groundwork for addressing some of the unaddressed questions and could support NIH efforts to track new data elements within the context of current administrative data collections (e.g., IMPAC II) or through new special data collections.
ORC Macro 6
1. AN OVERVIEW OF THE KIRSCHSTEIN-NRSA PROGRAM
The Ruth L. Kirschstein National Research Service Award (NRSA) program (Kirschstein-NRSA program) reached a 30-year milestone in 2005.3 Established by the National Research Service Award Act of 1974 (P.L. 93-348), the Kirschstein-NRSA program replaced all previous training authorities, consolidating them under a single program that provides predoctoral and postdoctoral research training. The overarching goal of the Kirschstein-NRSA program is to ensure that the U.S. biomedical and behavioral research labor force continues to be expanded and improved with a stream of new, well-trained investigators. The program was designed to direct National Institutes of Health (NIH) training funds toward developing investigators rather than health professionals and to discourage the use of this authority to support the clinical training of physicians. This program is the only current authority under which NIH supports the basic preparation of individuals for careers in biomedical research.4 This highly trained and diverse workforce will then be available to assume leadership roles in addressing the scientific research missions of NIH Institutes and Centers (ICs).5
Since its implementation in fiscal year (FY) 1975, the Kirschstein-NRSA program has focused on supporting predoctoral and postdoctoral training in a wide range of health-related disciplines where there is a need for personnel.6 Two kinds of awards have been made: 1) institutional training grants to public and nonprofit private institutions for multidisciplinary training programs or targeted training programs in specific research areas and 2) fellowships to individuals for supervised study with a senior scientist at an institution of the fellow’s choice. Both awards are conferred after competitive peer review. When NIH first awarded research training support in FY 1975, awards were made to 2,202 postdoctoral trainees and fellows; by FY 2002, 6,281 individuals received postdoctoral-level support.7
1.1. CHANGES IN KIRSCHSTEIN-NRSA POSTDOCTORAL POLICIES AND PRACTICES
In an effort to consolidate research training activities sponsored by NIH and the Alcohol, Drug Abuse, and Mental Health Administration, Congress enacted the National Research Service Award Act of 1974 to ensure that new training programs would be managed fairly and would be responsive to the needs of research. In the years since the Act was signed, the law governing the program has been modified several times in order to include new areas of research training, establish funding levels for selected disciplines and educational expenses, reduce the service obligation of Kirschstein-NRSA trainees, and encourage the recruitment of women and
3 Effective with the enactment of P.L. 107-206 in 2002, the National Research Service Award Program was renamed the Ruth L. Kirschstein National Research Service Award Program.4 NIH Task Forces for the Review of NIH Biomedical Research Training Programs. (1989). Review of the National Institutes of Health Biomedical Research Training Programs. Bethesda: NIH.5 Georgine Pion. (2001). The Early Career Progress of NRSA Predoctoral Fellows and Trainees, http://grants.nih.gov/training/career_progress/.6 The National Academy of Sciences, through its Committee on National Needs for Biomedical and Behavioral Scientists, has periodically issued reports specifying the number and type of trainees who should be funded under the Kirschstein-NRSA program.7 NIH Trainee and Fellow File (2004)
ORC Macro 7
underrepresented minorities for research careers. Congress also directed the National Academy of Sciences to provide periodic guidance on the fields in which researchers were likely to be needed and the numbers of individuals who would need to be trained. In 2005, the 12th such study was published.8
Kirschstein-NRSA policies, while attempting to secure diversity in health-related disciplines, institutions, and activities, are based on core values governing their operation and functioning. The awards follow a defined set of guidelines regarding participant eligibility (e.g., the program is limited to U.S. citizens and permanent residents), types of training supported, stipend levels, length of support, and appropriate use of funds. As with any program, policy changes have occurred since the program’s inception, most notably in 2001.9 Many of these changes have been made in order to improve the recruitment and retention of participants, streamline program administration, and address the need for biomedical and behavioral investigators in particular research areas.
1.1.1. Maximum Duration of Support
The duration of postdoctoral training varies with circumstances unique to the individual as well as broader influences such as the job market. For example, it has been suggested that individuals remain longer in postdoctoral positions when jobs are more difficult to find.10 It is the policy of the Kirschstein-NRSA program to limit the duration of support that an individual may receive through its various mechanisms.
Institutional T32 grants may be made for periods of up to 5 years and are renewable. The institution then appoints postdoctorates in varying numbers and for varying periods of support. No trainee may be appointed for less than 9 months during the initial period of appointment, except with the prior approval of NIH.11
No individual trainee may receive more than 3 years of support at the postdoctoral level, including any combination of support from institutional training grants and individual fellowship awards. Any extension of the total duration of trainee support at the postdoctoral level requires approval by the director (or an individual who has been delegated that authority by the director) of the NIH IC that supports the award. Requests for extensions must be made in writing by the trainee, endorsed by the director of the training program and the appropriate institutional official, and addressed to the director of the awarding component. The request must include a compelling justification for an extension of the statutory limits on the period of support.
8 Committee for Monitoring the Nation’s Changing Needs for Biomedical, Behavioral, and Clinical Personnel. (2005). Advancing the Nation’s Health Needs: NIH Research Training Programs. Washington, DC: National Academies Press. 9 In 2001, NIH amended the regulations governing NRSA to allow the incorporation of changes required by the enactment of the Alcohol, Drug Abuse, and Mental Health Administration Reorganization Act of 1992 (P.L. 102-231) and the National Institutes of Health Revitalization Act of 1993 (P.L. 103-43). 10 Howard H. Garrison and Susan A. Gerbi. (1998). “Education and Employment Patterns of U.S. Ph.D.’s in the Biomedical Sciences.” The FASEB Journal, 12:139–148.11 NIH National Research Service Award Institutional Research Training Grants (T32) Program announcements PA-02-109 (May 16, 2002) and PA-00-103 (June 1, 2000)
ORC Macro 8
1.1.2. Financial Support Package
Kirschstein-NRSA awards provide assistance to both individuals and the institutions sponsoring their training. The support package typically includes a stipend, tuition support, and an allowance to help defray other miscellaneous costs related to research training (e.g., supplies, a travel allowance to attend scientific meetings, health insurance). Through the early 1980s, NIH training grants and fellowships were generally able to cover the full costs of tuition and fees. By the early 1990s, however, a number of ICs began to cover only a portion of full tuition due to rapidly rising tuition costs. While this allowed the ICs to support a greater number of trainees and fellows, NIH announced a new standard in 1996 to reestablish uniformity of support: The Kirschstein-NRSA program would cover 100 percent of tuition and fees up to $2,000 and 60 percent of the costs above $2,000.12 A subsequent modification changed the formula to provide $3,000, plus 60 percent of the requested tuition in excess of $3,000.13 This formula is used to determine the tuition level provided via each competing grant; once established for a given competing grant, the same tuition level is used for noncompeting renewal awards during the project period.
Between FY 1998 and FY 2003, the NIH budget doubled, and increased funding for NRSA activities provided increases in trainee stipends and covered some of the escalating requests in the tuition category. Despite the return of more limited NRSA budget growth, requests and outlays for tuition continued to rise substantially in FY 2004 and 2005. As a result, NIH has frozen tuition expenses on competing renewals of T32 awards in FY 2006.
In 1975, there were six levels of postdoctoral stipends; by 1980, there were eight levels.14 In 1975, postdoctoral stipends ranged from $10,000 to $13,200, with an average of $11,750. By 2004, annual stipends ranged from $35,568 to $51,036, with an average of $43,775. In terms of 1980 dollars, the average 2004 stipend is about $15,500 when the NIH Biomedical Research and Development Price Index (BRDPI) is used as the deflator.
12 “Tuition Costs on NIH NRSA Training Grant and Fellowship Awards.” NIH Guide for Grants and Contracts, February 2, 1996.13 NIH Grants Policy Statement, Part II: Terms and Conditions of NIH Grant Awards, Subpart B: Terms and Conditions for Specific Types of Grants, Grantees, and Activities, December 200314 Stipend levels are based on years of postdoctoral experience and range from level 0 to level 7.
ORC Macro 9
Figure 1.1. Average Kirschstein-NRSA Trainee and Fellow Stipends, in Current and Constant Dollars (Base Year = 1980), FY 1975–2005
Note: Stipend amounts were adjusted to 1980 dollars using the BRDPI.Source: Dr. Walter T. Schaffer, NIH Office of Extramural Research, July 2004
Since 1980, the NIH training budget increased from about $196 million to $681 million, although in constant 1990 (1980) dollars, FY 2003 expenditures for training were just over $230 ($305) million.15 The NIH training budget declined from 8 percent of extramural research grant funding in FY 1981 to 4 percent in FY 1998. The training budget leveled off at about 4 percent in FY 2003.
Between FY 1998 and FY 2003, the training budget increased by about $253 million (59 percent) in current dollars. This growth resulted in part from an increase of 500 postdoctoral awards (and 1,000 predoctoral awards) and a significant increase in Kirschstein-NRSA stipends. These increases were more consistent with the total NIH budget, which doubled during those 5 years, from $13.6 billion to $27.2 billion.
1.1.3. Service Payback Requirement
The service payback requirement was established at the beginning of the Kirschstein-NRSA program and obligated training recipients to spend a year engaging in health research or teaching at a level averaging more than 20 hours per week of a full work year for each year of support received. In 1978, Congress waived the service payback requirement for short-term trainees who
15 Constant dollars were calculated using Consumer Price Index statistics published by the Department of Labor.
ORC Macro 10
participated in the Kirschstein-NRSA program for up to 3 months.16 The NIH Revitalization Act of 1993 specified that Kirschstein-NRSA recipients incur a service payback obligation only during their first 12 months of postdoctoral support.17 The Act further specified that the second and subsequent years of postdoctoral Kirschstein-NRSA training will serve to pay back the postdoctoral service obligation and included research training among payback activities for the first time. That is, postdoctoral trainees who continue under the Kirschstein-NRSA award for 2 years have fulfilled the payback obligation incurred during the first 12 months of support by the end of the second year.
Recipients with service obligations must begin to provide acceptable payback service on a continuous basis within 2 years of termination of Kirschstein-NRSA support. Requests for an extension specifying the need for additional time and the length of the required extension may be submitted to NIH. Those who do not pay back their obligation through continued Kirschstein-NRSA-supported training can satisfy it by serving in a position in which health-related research, research training, or teaching are the primary activities. For individuals who fail to fulfill their obligation through service, the United States is entitled to recover the total amount of Kirschstein-NRSA funds paid to the individual, with interest at a rate determined by the Secretary of the U.S. Department of the Treasury. Financial payback must be completed within 3 years, beginning on the date that the United States becomes entitled to recover the funds.18
1.2. GROWTH OF POSTDOCTORAL TRAINING PROGRAMS
Over the past 30 years, the national postdoctoral training enterprise in science and engineering has grown considerably, from nearly 17,000 postdoctorates in 1975 to 45,000 in 2002.19 For the traditional Kirschstein-NRSA postdoctoral training fields,20 the greatest average annual increase occurred in medical and other health fields, which showed an average annual increase of 5 percent. This was followed by the biological sciences, with an average annual increase of 4 percent, and psychology and other social sciences, which had an average annual increase of 2 percent.
Other fields of study grew at even faster annual rates, although the total number of individuals in those fields was not as large. For example, the average annual increase in the number of postdoctorates in mathematical and computer sciences was 7 percent between 1975 and 2002; engineering postdoctorates increased at an average rate of 5 percent per year.21
16 Community Mental Health Centers Extension Act of 1978. P.L. 95-622, Title II, Part D, 95th Congress, 2nd session, May 15, 1978. 17 National Institutes of Health Revitalization Act of 1993. P.L. 103-43, Title XVI, Section 1602. 103rd Congress, 1st session, June 10, 1993.18 Source: NIH PA-03-067, February 6, 2003, http://grants.nih.gov/grants/guide/pa-files/PA-03-067.html. Policies regarding the Kirschstein-NRSA payback obligation are explained in the Kirschstein-NRSA section of the NIH Grants Policy Statement, available at http://grants.nih.gov/grants/policy/policy.htm.19 National Science Foundation-NIH Survey of Graduate Students and Postdoctorates in Science and Engineering (2002)20 Biological sciences, medical and other health fields, and psychology and social sciences21 In this calculation, bioengineering was included with biological sciences rather than engineering.
ORC Macro 11
Figure 1.2. Number of Postdoctorates in Doctorate-Granting Colleges and Universities, by Major Field of Science and Engineering and Year, Fall 1975–2002
Note: No data were available for fall 1978.Source: National Science Foundation (NSF)-NIH Survey of Graduate Students and Postdoctorates in Science and Engineering, 2002
In 2002, nearly 40 percent of all postdoctorates (not just those supported by the Kirschstein-NRSA program) were in biological science departments, a proportion that has remained at about that level since 1979. Another 29 percent of all postdoctorates were in medical sciences and other health fields, a group that has increased about 5 percent since 1990. Less than 3 percent were in psychology and other social science departments.22
1.3. KIRSCHSTEIN-NRSA POSTDOCTORAL TRAINING MECHANISMS
Postdoctoral research provides new research and professional doctorates with an opportunity to expand their expertise and explore areas outside their doctoral fields. Section 487 of the Public Health Service Act (42 U.S.C. 288) provides authority for NIH to award Kirschstein-NRSA individual fellowships to support the predoctoral and postdoctoral training of individuals to undertake biomedical, behavioral, or clinical research at domestic and foreign, public and private institutions (profit and nonprofit). All NIH ICs except the Fogarty International Center (FIC) and National Library of Medicine (NLM) award Kirschstein-NRSA individual fellowships and institutional research training grants. FIC and NLM have unique funding authorities for fellowships and training grants that are separate from the Kirschstein-NRSA authority.23 This study focuses on the following Kirschstein-NRSA postdoctoral training programs: 22 NSF-NIH Survey of Graduate Students and Postdoctorates in Science and Engineering (2002)
ORC Macro 12
Ruth L. Kirschstein NRSA Institutional Research Training Grants (T32) Ruth L. Kirschstein NRSA Individual Postdoctoral Fellows (F32) Intramural NRSA Individual Postdoctoral Program (F35)
These training mechanisms are described in detail below.
1.3.1. T32 Traineeship Program
Institutional awards are funded through Ruth L. Kirschstein NRSA Institutional Research Training Grants (T32).24 Eligible institutions are awarded T32 grants to develop or enhance postdoctoral research training opportunities for individuals who are selected by the institution and who are training for careers in specific areas of behavioral and biomedical research. Only U.S. nonprofit private or public institutions can apply for these grants. The applicant institution must have a strong research program in the areas proposed for research training and must have the requisite staff and facilities to carry out the proposed program. In addition to providing postdoctoral research training for research doctorates, T32 grants provide a mechanism for the postdoctoral training of physicians and other health professionals who have otherwise limited research experience. Individuals receiving traineeship appointments must hold a Ph.D., D.V.M, D.D.S., M.D., or comparable doctoral degree from an accredited U.S. or foreign institution. Individuals with a professional doctorate can participate in the Kirschstein-NRSA postdoctoral T32 program as part of a research doctoral degree program. Postdoctorates are encouraged to participate in at least 2 years of research, research training, or other comparable activities beginning with their first traineeship appointment. It has been shown that there is a strong correlation between duration of training and retention in post-training research activity.25
1.3.2. F32 and F35 Fellowship Programs
Individual awards are funded through the Ruth L. Kirschstein NRSA Individual Postdoctoral Fellows (F32) program.26 Before submitting a fellowship application, an applicant must identify a sponsoring institution and an individual who will serve as a sponsor and supervise the training and research experience. The F32 fellowships provide research training to postdoctorates who have received a Ph.D., M.D., D.O., D.C., D.D.S., D.V.M., O.D., D.P.M., Sc.D., Eng.D., D.P.H.,
23 NIH Grants Policy Statement, Part II: Terms and Conditions of NIH Grant Awards, Subpart B: Terms and Conditions for Specific Types of Grants, Grantees, and Activities, December 200324 Institutional awards include the Ruth L. Kirschstein NRSA Short-Term Institutional Research Training Grants (T35), which are designed primarily to provide research training experience for students in health professional schools during summer breaks. The purpose of the short-term appointments is to provide opportunities for participation in biomedical or behavioral research in an effort to attract students to research careers. T35 grants may also be used to support other types of postdoctoral training in emerging scientific areas related to the mission of a particular IC. The T35 training mechanism was not included in this study.25 Committee for Monitoring the Nation’s Changing Needs for Biomedical, Behavioral, and Clinical Personnel. (2005). Advancing the Nation’s Health Needs: NIH Research Training Programs. National Academies Press: Washington, DC.26 Individual awards support Ruth L. Kirschstein NRSA Senior Fellows (F33). These fellowships allow experienced scientists to make major changes in the direction of their research careers, broaden their scientific background, or acquire new research training. The F33 training mechanism was not included in this study.
ORC Macro 13
D.N.S., N.D., Pharm.D., D.S.W., Psy.D., or equivalent degree in order to broaden their scientific background and further their potential for research in health-related areas. Individuals may receive up to 3 years of aggregate Kirschstein-NRSA support at the postdoctoral level, including any combination of support from Kirschstein-NRSA institutional traineeship and individual fellowship awards.
Between FY 1979, when the last of the remaining training commitments under the pre-Kirschstein-NRSA training authority were funded, and FY 2002, the average annual increase in the number of individuals supported was only 0.4 percent (see figure 1.3). The number of Kirschstein-NRSA postdoctorates peaked at 6,710 in FY 1980 and reached a low of 5,757 in FY 1989. The number of trainees and fellows in the program has remained in the 6,000–6,500 range since 1990.
The intramural F35 award provides individual physicians, dentists, and veterinarians who have research experience with an opportunity to prepare for research careers in an NIH laboratory as part of the NIH intramural program.27 They were last awarded in FY 1998.
27 http://www.niaid.nih.gov/ncn/glossary/default4.htm#intramural
ORC Macro 14
Figure 1.3. Number of Individuals Receiving Kirschstein-NRSAPostdoctoral Support per Year, FY 1975–2002
Note: Kirschstein-NRSA postdoctoral funding mechanisms include T32, T35, F32, F33, and F35. The data presented in figure 1.3 represent the number of individuals supported each year.Source: NIH Trainee and Fellow File (2004)
1.4. MAJOR STUDY QUESTIONS AND METHODOLOGY
The major purpose of this report is to examine and assess the role of Kirschstein-NRSA postdoctoral training support in the career outcomes of researchers in the biomedical and behavioral sciences. The most traditional outcome measures for such an assessment, which are the focus of this study, are:
Success in obtaining funding for research through various NIH grant mechanisms Publication of research papers in peer-reviewed journals Employment in a position that provides the individual with opportunities to add to overall
scientific knowledge in the biomedical and behavioral sciences
The linkage between Kirschstein-NRSA postdoctoral support and these outcomes rests on the following assumptions:
1. Receiving Kirschstein-NRSA postdoctoral support enables enhanced training, offers additional research opportunities, and promotes development of professional relationships that will assist individuals in their early career development.
ORC Macro 15
2. Kirschstein-NRSA postdoctoral support, particularly for biomedical researchers, confers advantages beyond those available from other kinds or sources of postdoctoral support.
3. The effects of Kirschstein-NRSA postdoctoral support can be separated from the effects associated with participating in predoctoral programs and from other demographic and situational characteristics.
We used several study groups to examine the first two assumptions. Under the first assumption, we focused on exploring the differences between Kirschstein-NRSA postdoctorates and the population of researchers in general. Any differences were expected to favor Kirschstein-NRSA postdoctorates for a number of reasons, including the fact that these programs (particularly in the biomedical sciences) have become an important step in preparing individuals for research careers. Other individuals who earn research doctorates may have chosen to directly pursue employment conducting research in the private sector or may not have demonstrated the level of research career focus expected of most Kirschstein-NRSA postdoctorates. Some individuals without Kirschstein-NRSA postdoctoral support will achieve successful research career outcomes, but we expected that as a group, they would not achieve successful research career outcomes to the extent that those supported by the Kirschstein-NRSA program would.
Under the second assumption, we explored the differences between Kirschstein-NRSA postdoctorates and those individuals who received other postdoctoral research training support. We expected to be less likely to find differences favoring Kirschstein-NRSA postdoctorates, in part because many non-Kirschstein-NRSA postdoctorates received training in the same institutional environment as those who received Kirschstein-NRSA support. Individuals supported with funding from other Federal and non-Federal training sources may have received an equivalent level of training and career opportunities.28
The third assumption is critical for making any assessment of the specific effects of the Kirschstein-NRSA postdoctoral program. To assess the effects of Kirschstein-NRSA postdoctoral research training support, we have attempted to control for a variety of characteristics that relate to the kind of training and environment individuals were exposed to as predoctorates, the career interests of these individuals, the family and personal background of individuals who make certain career choices, and the capabilities of these individuals.
A conceptual framework for capturing these assumptions is presented in figure 1.4. This framework suggests a model where predoctoral experiences, demographics, and opportunities create and/or allow an interest in pursuing postdoctoral study. In other words, participation in postdoctoral studies is partially due to a self-selection process in which individuals express interest in pursuing research training as a first step in becoming an independent investigator, and partially due to an institutional selection process in which potential researchers are evaluated on their potential for contributing to the scientific field and placed into situations that would allow them to build on that potential. These selection factors could, by funneling individuals with greater interest in pursuing scientific research through postdoctoral studies, suggest that predoctoral experiences have a major effect on eventual research career outcomes.
28 NSF and other Federal agencies and private organizations, such as the Howard Hughes Medical Institute with its endowment of $12.8 billion (FY 2004), play a significant role in supporting postdoctorates in the biomedical and behavioral sciences.
ORC Macro 16
Figure 1.4 also suggests a number of other comparisons that are useful for understanding the effects of various Kirschstein-NRSA postdoctoral support mechanisms, including:
Successful versus unsuccessful Kirschstein-NRSA fellowship applicants. This comparison examines those who applied for an F32 or F35 fellowship award and either succeeded or failed on the basis of NIH peer review of their application(s). Applications are evaluated based on the 1) research topic, 2) ability of the individual to conduct the research, 3) ability of the sponsoring institution to support the research, and 4) relationship with a mentor. To the extent that the application accurately represents the capabilities of the individual to produce quality research, we might also expect the success or failure of the application to be associated with future research success.
The successful postdoctoral fellowship group contains individuals who received a fellowship at any time within the period of this study (1975–1992), even if they failed at their first attempt to obtain a fellowship. Unsuccessful applicants are individuals who applied at least once and never received an F32 or F35 award. We excluded individuals from the unsuccessful applicant group who failed to obtain Kirschstein-NRSA fellowship funding but received a traineeship.
Kirschstein-NRSA postdoctoral fellows versus Kirschstein-NRSA postdoctoral trainees. Because fellowships are individually competed and allow successful individuals to pursue their own research interests, they are similar in some respects to a competing research grant. As a result, a fellowship may advance the career of the fellowship recipient to a greater extent than Kirschstein-NRSA traineeship support.
Comparison of Kirschstein-NRSA fellows and trainees by months of support. The Kirschstein-NRSA program provides support to individuals for up to 36 months, and possibly longer with the consent of the supporting NIH IC. However, there is considerable variation in the duration of support, thus leading to an inquiry into how length of Kirschstein-NRSA support affects outcomes.
ORC Macro 17
Figure 1.4. Conceptual Framework for Examining the Relationship of Kirschstein-NRSA Postdoctoral Training and Research Outcomes
1.4.1. Study Populations
The study and comparison groups were constructed to include:
Doctorates who received Kirschstein-NRSA postdoctoral support (T32 traineeships, F32 and F35 fellows). This group is the primary study group. Because NIH maintains data on these individuals, statistics for the entire population can be generated on many measures used in this study. A subsample of individuals in this study group who responded to the 1995 Survey of Doctorate Recipients (SDR) was used to examine publication and employment outcomes.
Doctorates who were unsuccessful Kirschstein-NRSA postdoctoral fellowship applicants. This is another group for which NIH maintains data, and therefore all individuals who applied for a fellowship can be identified and examined regarding certain outcomes. As with successful fellows, a subsample of these individuals who responded to the 1995 SDR was used to examine publication and employment outcomes.
Research doctorates with non-NIH postdoctoral training support. The only source for identifying this group is the SDR, since information on postdoctoral support is collected by
ORC Macro 18
Characteristics of Predoctoral Study Demographics
Interest in Pursuing Postdoctoral Studies
ReceivedKirschstein-
NRSA Fellowship(F32, F35)
Appointed toKirschstein-
NRSA Traineeship
(T32)
ReceivedOther
Postdoctoral Support
No Evidenceof Postdoctoral
Support
Funding Outcomes
Publication Outcomes
Employment Outcomes
Outcomes
that survey. This group included only individuals who responded to the 1995 SDR.
Research doctorates with no postdoctoral training support. As with the previous group, this group was identified by responses to the 1995 SDR.
Within the first two study groups we can make a distinction between professional doctorates (e.g., M.D.’s) and research doctorates (e.g., Ph.D.’s). For the latter two groups, we do not have information on those who only have a professional degree, since the SDR focuses on individuals with research doctorates.29 These groups were developed to establish a set of study groups comparable with each other.
1.4.1.1. Analysis of Professional Doctorates
A very large proportion of Kirschstein-NRSA postdoctoral recipients have M.D.’s or other professional degrees. Some of them also have research doctorates. Since Kirschstein-NRSA postdoctoral support is focused on developing researchers, individuals with M.D.’s and other health professionals receiving support are assumed to have an interest in pursuing research careers and can therefore be assessed by the three types of outcomes (i.e., NIH research funding, publication, and employment) specified in the conceptual framework. The only major sources of information available for M.D.’s and other professional doctorates were the NIH Consolidated Grant Applicant File (CGAF) and Trainee and Fellow File (TFF).30 Thus, we could readily describe the postdoctoral experiences of M.D.’s in terms of their NIH research training and success in obtaining NIH research funding, but we had less information on their predoctoral experiences, their expressed interest in research, and their publications and employment history.
1.4.2. Selection of Study Populations
The research methodology used information contained in the CGAF and TFF data files as well as the Doctorate Records File (DRF) and SDR. The selection of the study population was keyed to the availability of outcomes data from these sources. It should be noted that the timeframe for the study included individuals who received support from FY 1975, when the Kirschstein-NRSA program started, through FY 1992. FY 1992 was selected as the last year in order to provide at least 10 years after the completion of the postdoctorate in which to observe outcomes.
1.4.2.1. Selection of the Kirschstein-NRSA Study Population
In all, 27,832 research doctorates received some Kirschstein-NRSA postdoctoral support between FY 1975 and FY 1992. This population was further defined by the following criteria:
29 Groups used in the Garrison and Brown study, based on postdoctoral plans, were not explicitly examined here except as covariates in the analysis.30 The major sources of predoctoral and outcomes information for this study are the Survey of Earned Doctorates (Doctorate Records File) and SDR. Both surveys target individuals with research doctorates; those with professional degrees are excluded.
ORC Macro 19
No NIH support prior to the first Kirschstein-NRSA postdoctoral support31 No Kirschstein-NRSA postdoctoral support after FY 1992 First postdoctoral Kirschstein-NRSA award/appointment not more than 11 months prior to
the receipt of the research doctorate Research doctorates in traditional biomedical and behavioral science fields Doctorate awarded no earlier than 1965
The population as defined by these criteria contained 23,696 individuals. Information from the CGAF and TFF was used to examine selected predoctoral characteristics, demographics at the time of degree award, Kirschstein-NRSA postdoctoral award characteristics, and NIH research grant outcomes.
In addition, some 6,891 professional doctorates (without a research doctorate) were selected for certain analyses. These individuals met the same basic requirements used to select the research doctorate population. The study explored the predoctoral characteristics of the population, their Kirschstein-NRSA training support characteristics, and outcomes related to NIH research funding. However, it did not include information on publication and employment outcomes.
1.4.2.2. Kirschstein-NRSA and Other Study Group Subpopulations—SDR Analyses
We used SDR data to answer questions related to employment outcomes.32 A subpopulation of the Kirschstein-NRSA study group and two comparison groups were formed from individuals who responded to the 1995 SDR. The three groups included:
All individuals identified in the Kirschstein-NRSA study group who responded to the 1995 SDR
U.S. citizens or permanent residents who responded to the 1995 SDR, received at least one postdoctoral appointment but did not receive NIH postdoctoral support, and otherwise met the criteria specified for the Kirschstein-NRSA study group
U.S. citizens or permanent residents who responded to the 1995 SDR, indicated that they had no postdoctoral appointments, and received their research doctorates between 1965 and 1992 in the same fields as the Kirschstein-NRSA study group
Thomson ISI Web of Science provided publication and citation information for a subsample of individuals selected from the SDR population. The exact nature of the ISI subsample is described in chapter 4.
31 The Kirschstein-NRSA training program began in 1975; prior to that, individuals could have received support from other NIH vehicles. Such individuals (1,317 who received their doctorates between 1965 and 1992) were eliminated from the analysis.32 The SDR, in specific years, also collected information on publications. We used this source for a subsample of individuals.
ORC Macro 20
1.4.3. Data Files and Data Elements
This study was initiated as a comprehensive examination of the Kirschstein-NRSA postdoctoral research training program using data from the following sources:
CGAF—This file contains the history of individual NIH grant applications and awards.
TFF—This file contains the histories of individuals’ interactions with the Kirschstein-NRSA training programs. It includes anyone who applied for Kirschstein-NRSA fellowship or received a traineeship appointment.
DRF—This file contains the results of the Survey of Earned Doctorates (SED), which collects information on research doctorates who earned degrees at accredited U.S. institutions. It is useful for describing characteristics relating to predoctoral studies and characteristics of the individual at the time of the receipt of the doctorate.
SDR—This is a longitudinal, biennial survey of a stratified sample of research doctorates. During each survey cycle, a sample of the newest two cohorts of research doctorates is added to the survey frame, while the two oldest cohorts are eliminated. In all, 42 cohorts are represented in each biennial survey. The information collected is aimed at updating information on employment, professional development and accomplishments, additional training, and personal characteristics.
ISI Publications Data—These data represent records of scientific publications authored by a subsample of SDR recipients who were included in the study. Publication and citation information from 1990 to 2001 is represented.
1.5. FOCUS OF THE STUDY
This study focuses on the effect of Kirschstein-NRSA postdoctoral support on research career outcomes and explores evidence that such support produces individuals who have productive, research-focused careers. In our report, we examine the characteristics of former Kirschstein-NRSA postdoctoral trainees and fellows and their early career progress in biomedical and behavioral research. This can be considered a sequel to the 1986 assessment of the Kirschstein-NRSA program.33
The remaining chapters in this report present the results of our analyses. In chapter 2, we explore the backgrounds of research doctorates in the Kirschstein-NRSA postdoctoral study group and make comparisons with research doctorates who did not receive such support. Because we had sufficient data for the entire population of individuals with records in the CGAF, TFF, and DRF, we used these sources for the analyses. We focused on simple differences in exploring the backgrounds of research doctorates in the biomedical and
33 Howard H. Garrison and Prudence W. Brown. (1986). The Career Achievements of NIH Postdoctoral Trainees & Fellows. NIH Program Evaluation Report, Committee on National Needs for Biomedical and Behavioral Research Personnel, Institute of Medicine. Washington, DC: National Academies Press.
ORC Macro 21
behavioral sciences. These differences provide us with information for understanding the pre-Kirschstein-NRSA postdoctoral support factors that might affect career outcomes.
In chapter 3, we further analyze the Kirschstein-NRSA postdoctorates in order to understand the potential effects of various training mechanisms sponsored by this program. In this chapter, we focus on differences among individuals who are involved in different Kirschstein-NRSA postdoctoral research training programs.
In chapter 4, we present data describing the three major outcome measures—employment, publications, and research grants, focusing on outcomes during specified intervals after the award of the doctorate. For example, success in obtaining NIH research funding through major grant mechanisms was limited to 10 years after the doctorate was awarded. The analyses use multivariate techniques to control for differences in characteristics of predoctoral training, demographics, and other factors.
In chapter 5, we present a summary of our findings and conclusions.
ORC Macro 22
2. PREDOCTORAL EXPERIENCES AND BACKGROUND CHARACTERISTICS OF NEW RESEARCH DOCTORATES
The importance of the predoctoral experiences of individuals in the behavioral and biomedical sciences for predicting later success as research scientists has been examined in studies by Garrison and Brown34 and Pion.35 In the study by Garrison and Brown, the prestige of the doctoral institution and an individual’s performance during his or her predoctoral studies were found to be positively related to research-related outcomes such as employment, publications, and success in procuring NIH research funding. In an analysis focusing on predoctoral students, Pion found that those with Kirschstein-NRSA predoctoral support were more likely to succeed on a number of research productivity measures than individuals without such funding. In attempting to measure outcomes, these studies also examined the effects of demographic characteristics, such as gender, and other background characteristics, such as marital status and number of dependents at the time the doctorate is awarded. The underlying concept is that these factors can facilitate success in achieving both postdoctoral support and independent research careers. Alternatively, these factors may direct individuals away from considering postdoctoral studies. This chapter presents information for understanding the differences between those who received Kirschstein-NRSA postdoctoral support and those who did not, as well as the differences among Kirschstein-NRSA postdoctoral recipients with different kinds of support.
2.1. CONSIDERATIONS IN CHARACTERIZING RESEARCH DOCTORATE BACKGROUNDS
The Garrison and Brown study identified a number of characteristics related to predoctoral backgrounds. In particular, they examined background characteristics related to the following:
Prestige of institutions from which individuals received their research doctorates. Prestige, for Garrison and Brown, was measured by an institution’s scholarly achievements in various biomedical disciplines.36 Such prestige measures support the notion that the particular environment in which doctoral studies are pursued provides additional benefits beyond those associated with the research doctorate. Positive factors in this environment could include the reputation of the research faculty and the kinds of research opportunities that exist.37 For this analysis, we will examine the research intensity of the institutions from which individuals received their doctorates, with the idea that research intensity provides an overall indication of research opportunities, quality of the faculty, and emphasis on research—all of which are expected to be related to the pursuit of postdoctoral studies and success in research-related career outcomes.
34 Garrison and Brown, 1986.35 Pion, Georgine, 2001.36 The Garrison and Brown study used results from the 1982 Assessment of Research-Doctorate Programs, for which an average for each university was computed from the mean ratings of scholarly quality of faculty in biochemistry, cellular/molecular biology, microbiology, and physiology programs. 37 It should be noted that these factors can vary considerably within an institution, with certain disciplines or departments providing better opportunities than others. Thus, this effect may be less profound when examined in an institutional context than when examined in a departmental context.
ORC Macro 23
Kirschstein-NRSA support during predoctoral studies. The Pion study found that individuals with predoctoral Kirschstein-NRSA support were more likely to succeed in achieving certain outcomes tied to a research career. We will explore whether individuals receiving Kirschstein-NRSA postdoctoral support were more likely than others to receive Kirschstein-NRSA predoctoral support, thus establishing a linkage that would more likely result in a research career.
Time spent in pursuit of the doctorate. The time spent in pursuit of the doctorate may, as inferred by the Pion study, be a success measure. There are various reasons why predoctoral students spend more or less time in pursuit of the doctorate, but we might expect that those who finish sooner are more directed and organized, which may indicate a strong aptitude for continued postdoctoral study and research.
Gender and marital status characteristics. These characteristics have been critical for describing scientific labor market participation rates. We would expect gender to be less of a factor than in the past because of the reduction in gender stereotyping and increased opportunities for women to pursue scientific careers. Marital status and number of dependents, however, may still influence career decisions because they impose a financial burden on individuals that may reduce the attraction of a postdoctoral position, compared with a higher-paying position.
Postdoctoral plans of the doctorate recipients. Stated plans upon receipt of the doctorate,
keeping factors relating to quality of training and abilities constant, might be expected to be a very strong predictor of actual postdoctoral involvement and possible career aspirations. It is certainly an indicator of an underlying interest in pursuing a research career.
Each of these factors was explored in previous studies and relates to a characteristic that may influence the postdoctoral careers of new doctorates. Some characteristics involve choices made by the individual (e.g., plans to pursue a Kirschstein-NRSA fellowship), while other characteristics are potential selection factors that will persuade NIH to provide an award to a postdoctoral fellowship applicant or an institution to provide the individual with a postdoctoral traineeship. We will focus on the same or a similar measure in our examination.
2.2. DATA SOURCES AND DEFINITIONS
Data presented in this chapter derive from the DRF, which contains the results of the annual SED. The CGAF and TFF were used to differentiate among individuals by the type of postdoctoral support received. It should be noted that the SED focuses on collecting information from individuals receiving a research doctorate. Individuals who received only an M.D. are not included in the DRF38 and are therefore not included in the analysis in this chapter.
With regard to describing recipients of research doctorates in terms of the characteristics presented in section 2.1, comparisons among various groups are made based on their
38 Individuals who have dual M.D. and research doctoral degrees are surveyed at the time they receive their research doctorates. Information about an M.D. degree awarded prior to the research doctorate is included in the DRF.
ORC Macro 24
Kirschstein-NRSA postdoctoral status. Table 2.1 describes several mutually exclusive study groups that were used for the background analysis.
Table 2.1. Primary Study Groups Used for the Analysis
Group Description Number in Analysis
Kirschstein-NRSA fellows
Research doctorates who received only Kirschstein-NRSA postdoctoral fellowship (F32 and F35) support. They received their first and last Kirschstein-NRSA fellowship between 1975 and 1992.
7,720
Kirschstein-NRSA trainees
Research doctorates who received only Kirschstein-NRSA postdoctoral traineeship (T32) support. They received their first and last Kirschstein-NRSA traineeship between 1975 and 1992.
13,894
Kirschstein-NRSA trainee/fellows
Research doctorates who received both Kirschstein-NRSA postdoctoral traineeship (T32) and fellowship (F32 and F35) support. They received their first and last Kirschstein-NRSA support between 1975 and 1992.
2,355
Unsuccessful fellowship applicants
Research doctorates who applied for but did not receive a Kirschstein-NRSA fellowship and did not receive Kirschstein-NRSA traineeship support between 1975 and 1992.39 The individuals in this group may have received postdoctoral support from other NIH mechanisms or other agencies during this period, or they may not have received any postdoctoral support.
9,725
Other research doctorates
Research doctorates in the biological or behavioral sciences who did not apply for Kirschstein-NRSA postdoctoral fellowships and did not receive Kirschstein-NRSA traineeship support of any type between 1975 and 1992. These individuals were identified from the SDR, which follows cohorts of individuals identified by the year in which they received their doctorates.40 They would include individuals performing postdoctoral work but do not include individuals from any of the above groups. This group was drawn from the 1995 SDR survey.
8,164
Sources: CGAF, TFF, DRF, SDR
The analysis includes individuals who received their research doctorates between 1965 and 1992. This means that there are some individuals in the population who received their first doctorates between 1965 and 1974 but received their first Kirschstein-NRSA postdoctoral support in 1975 or later. This group included 466 fellows, 755 trainees, and an additional 58 individuals who received both traineeship and fellowship support during their postdoctoral studies.
The first four groups include all individuals who met the above criteria and therefore represent the entire population of Kirschstein-NRSA postdoctorate recipients.41 The last group (other research doctorates) consists of individuals drawn from SDR respondents and therefore constitutes a sample. Results for this sample are subject to sampling error, although this is somewhat ameliorated by the large sample size. In presenting results in this chapter, we used the 1995 sampling weights given by the SDR to determine percentages.42 Given that we are essentially dealing with censuses or very large samples, almost any percentage point difference is 39 This group may include individuals who received support after 1992 and may include individuals who received fellowship support from sources other than the F32, F35, and T32 mechanisms.40 See section 1.4.3 for more information about the SDR. 41 Because the Kirschstein-NRSA groups, including unsuccessful fellowship applicants, represent the population, sampling error is not an issue and any difference can be considered a significant difference. 42 The 1995 sampling weights reflect the number of individuals in the population during 1995 represented by each respondent. The weighted and unweighted proportions for the cross-tabulations presented in this chapter are very similar. In certain tables where we present frequencies, we computed these numbers according to the total number of respondents and not the population totals.
ORC Macro 25
Kirschstein-NRSA postdoctoral support recipients were more likely (by a 9-percent margin) to receive their research doctorates from institutions identified by the Carnegie Classification system as being research focused.
Kirschstein-NRSA postdoctoral support recipients were more likely (by a 9-percent margin) to have received their research doctorates from one of the top 100 institutions ranked by NIH research funding.
While there are no differences between Kirschstein-NRSA fellows and trainees who have10–12 months of support, fellows who have 25–36 months of support tend to come from more research-intensive institutions (more so than trainees).
a “significant” difference.43 The issue then is how to identify important differences. In the following analyses, we generally do not consider differences of less than 5 percentage points to be notable.44
These comparison groups allow us to explore various relationships related to the predoctoral background of research doctorates as defined in section 2.1. In this section, we ask “How do these groups differ,” not “What factors are important in determining the type of Kirschstein-NRSA postdoctoral support received or whether any Kirschstein-NRSA postdoctoral support is received at all.”
2.3. DO KIRSCHSTEIN-NRSA POSTDOCTORAL SUPPORT RECIPIENTS RECEIVE THEIR RESEARCH DOCTORATES FROM MORE RESEARCH-INTENSIVE INSTITUTIONS?
Following the lead of previous studies, we hypothesized that predoctoral research opportunities and the chance to work with recognized investigators are important factors in determining an individual’s postdoctoral career trajectory. These factors are more likely to direct individuals to prestigious or interesting research venues during their postdoctoral studies. In this section, we examine the research emphasis of the predoctoral institution under the assumption that this emphasis will provide the individual with a greater range of career options, including the pursuit of postdoctoral studies. The institution’s involvement in research is indicated by its Carnegie Classification of Institutions of Higher Education and its ranking in research funding from NIH.
43 Differences in the Kirschstein-NRSA populations would not reflect sampling errors and therefore are “real” differences. However, differences may be due to measurement error.44 It must be emphasized that this chapter is purely descriptive in nature and that the 5-percentage-point difference was selected to illustrate notable differences or differences that might distinguish among study groups and would provide some insight into what group delineations will be useful for the outcomes analysis in chapter 4. Although there are a number of reasons for avoiding discussions of statistical power and effective sample size, it might be useful to indicate that if we were designing this study and selecting samples, the number of observations used in this chapter provides enough power to detect differences in proportions of 3 percent at power = .80 and alpha = .05, given the most restrictive requirements in base proportions being compared.
ORC Macro 26
2.3.1. Level of Research Intensity as Measured by the Carnegie Classification
Figure 2.1 provides information on the various comparison groups by Carnegie Classification. The classifications presented in this chapter, as defined in 1987,45 include:
Research I university—These institutions offer a full range of baccalaureate programs, are committed to graduate education through the doctoral degree, and give high priority to research. Annually they receive at least $33.5 million in Federal support for research and development and award at least 50 Ph.D. degrees.
Research II university—These institutions offer a full range of baccalaureate programs, are committed to graduate education through the doctoral degree, and give high priority to research. Annually they receive between $12.5 and $33.5 million in Federal support for research and development and award at least 50 Ph.D. degrees.
Other doctorate-granting universities—Any other university not included in the preceding classifications
For the purposes of the analysis, we focused on the representation of individuals receiving their degrees from Research I and Research II universities relative to the other doctorate-granting universities. As the definitions indicate, Research I and Research II universities are the most research-intensive institutions, and we expect that most individuals with an interest in pursuing research would emerge from them. The focus of this examination is whether there are differences in the research intensity of institutions that awarded degrees to Kirschstein-NRSA postdoctorates and institutions that awarded degrees to other research doctorates. Any differences may indicate that Kirschstein-NRSA postdoctorates may be better prepared to undertake the requirements of postdoctoral training and a subsequent research career.
As figure 2.1 shows, about 82 percent of all Kirschstein-NRSA postdoctorates received their degrees from institutions classified as Research I and II universities. This compares with 73 percent of other research doctorates who received their degrees from Research I and II universities. Thus, it seems that there are differences between Kirschstein-NRSA postdoctorates and other research doctorates in the research intensity of the institutions where they received their degrees, as defined by the Carnegie Classification.
In examining the fact that the training provided by research-intensive institutions may provide Kirschstein-NRSA postdoctorates with an additional advantage, we may also find that Kirschstein-NRSA fellows have an advantage over those who applied for and were not awarded a fellowship and over those who were appointed to a Kirschstein-NRSA traineeship.
Figure 2.1 also shows that:
45 The institutional classification at the time of receipt of the doctorate was used, and we therefore reverted to the classification system that was employed at that time. It should be noted that before 1987, there was a set quota for being included in the Research I (top 50 schools in Federal financial support) and Research II (top 100 schools in Federal financial support) categories. Currently there are two classifications: doctoral/research universities—extensive and doctoral/research universities—intensive.
ORC Macro 27
When postdoctoral fellows are compared with unsuccessful fellowship applicants, the respective percentages of those who received their doctorates from Research I and Research II institutions are 85 percent and 83 percent, respectively, a relatively small difference. Unsuccessful fellowship applicants (who also did not receive a traineeship) were slightly more likely to receive their doctorates from the highest-ranked research universities than Kirschstein-NRSA trainees, but again the difference was not notable.
Among Kirschstein-NRSA postdoctoral support recipients, 85 percent of those with Kirschstein-NRSA fellowship support received their doctorates from Research I and Research II universities. This compares with 81 percent for Kirschstein-NRSA trainees and 81 percent for those who had both Kirschstein-NRSA traineeship and fellowship support. We do not consider this to be a notable difference, based on our 5-percentage-point criterion.
When compared with other research doctorates, the data suggest that individuals applying for or receiving Kirschstein-NRSA postdoctoral support tend to do their predoctoral work at more research-intensive institutions as defined by the Carnegie Classification, but differences among these groups are not notable.
Figure 2.1. Percentage of Research Doctorates Receiving Their Degrees From Carnegie Classification Research I and Research II Universities,*
by Type and Receipt of Kirschstein-NRSA Postdoctoral Support
* Classifications are at the time of receipt of the doctorate.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.1.
Length of Kirschstein-NRSA postdoctoral support may be used to differentiate between fellows and trainees. The research intensity of the institutions granting the doctorate may affect the length of postdoctoral studies, and this may vary by type of postdoctoral support.
ORC Macro 28
When Kirschstein-NRSA postdoctoral support recipients are classified by the months of Kirschstein-NRSA postdoctoral support received, the patterns represented in figure 2.2 emerge. Each of the bars in this figure represents the percentage of individuals who attended Research I and II universities. The data show that for those individuals with at least 10–12 months of support, there is little difference between Kirschstein-NRSA postdoctoral fellows and trainees in the proportion of those receiving their degrees from Research I and II universities. However, if we consider those individuals with 25–36 months of support, there is a 7-percentage point difference between Kirschstein-NRSA postdoctoral fellows and trainees, suggesting that fellows who receive more months of postdoctoral support tend to receive their degrees from Research I and II institutions to a greater degree than trainees. Conversely, proportionally speaking, more trainees who received more months of postdoctoral support were awarded their research doctorates from institutions other than those classified as Research I or II.
This leads us to conclude that profiles of Kirschstein-NRSA postdoctoral fellows and trainees in terms of the research intensity of the institutions from which they received their doctorates are similar when the length of support is for shorter periods but differ when we consider those who have received more than 2 years of support. In short, the research intensity of the institution seems to be a factor for distinguishing Kirschstein-NRSA fellows from trainees with greater length of support and therefore validates both the use of research intensity and months of support in our analysis of outcomes.
ORC Macro 29
Figure 2.2. Percentage of Kirschstein-NRSA Fellows, Trainees, and Trainee/Fellows Receiving Their Research Doctorates From Carnegie Classification Research I and II
Universities,* by Months of Kirschstein-NRSA Postdoctoral Support
* Classifications are at the time of receipt of the doctorate.** Data for the trainee/fellow groups with 1–9 months and 10–12 months of Kirschstein-NRSA postdoctoral support are omitted because of the small number of observations.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.1.
2.3.2. Level of Research Intensity as Measured by NIH Research Funding
A second measure of research intensity is the level of NIH research support received by the institution from which the individual received his or her research doctorate. Institutions were ranked by the amount of NIH research grant support received during the year in which the doctorate was awarded. Our expectation was that Kirschstein-NRSA postdoctorates would be clustered in the top-ranking schools, compared with those who did not receive Kirschstein-NRSA funding.
Figure 2.3 presents the proportions of research doctorates with degrees from the top 100 institutions ranked by NIH extramural funding. It shows that:
Individuals receiving Kirschstein-NRSA postdoctoral support tended to have doctorates from institutions ranked in the top 100 in NIH funding to a greater degree than those not receiving such postdoctoral support. About 67 percent received their doctorates from institutions ranked in the top 100, compared with 58 percent of individuals who had no Kirschstein-NRSA postdoctoral support.
There is little difference between the proportions of those receiving and those unsuccessfully applying for Kirschstein-NRSA fellowships in terms of the research intensiveness of the institutions where they received their degrees.
ORC Macro 30
Marginally, trainees were more likely than fellows to have received their doctorates from institutions that received lesser amounts of NIH research funding, but the difference is only about 3 percentage points, not notable according to our criterion for identifying differences.
Results using NIH research funding support the analyses using the Carnegie Classification; that is, Kirschstein-NRSA postdoctorates and those applying for postdoctorates tend to have degrees from research-intensive institutions. Again, this analysis is investigating whether research intensiveness, as one measure reflecting the quality of research training received prior to postdoctoral studies, is useful for describing inherent differences between Kirschstein-NRSA postdoctorates and other research doctorates.
Figure 2.3. Percentage of Research Doctorates Receiving Their Degrees From the Top 100 Institutions Ranked by NIH Extramural Funding,*
by Type and Receipt of Kirschstein-NRSA Postdoctoral Support
* Ranking of institutions is at the time of receipt of the doctorate.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.2.
Figure 2.4 shows the percentage of Kirschstein-NRSA postdoctoral recipients who received their degrees from institutions ranked in the top 100 in NIH research funding classified by the number of months of Kirschstein-NRSA postdoctoral support received. The data indicate that, among individuals with 10–36 months of support, the percentage of fellows with degrees from institutions ranked in the top 100 increases as the number of months of Kirschstein-NRSA postdoctoral support increases, while for trainees the percentage decreases. Although the actual trends may not be notable, the differences between fellows and trainees increase from no difference to 5 percentage points, thus showing a potential divergence in the profiles of the two groups. These results are consistent with the analysis based on the Carnegie Classification and
ORC Macro 31
indicate that there are differences between Kirschstein-NRSA fellows and trainees who have had more postdoctoral support.
Figure 2.4. Percentage of Kirschstein-NRSA Fellows, Trainees, and Trainee/Fellows Receiving Their Research Doctorates From the Top 100 Institutions Ranked by
NIH Extramural Funding,* by Months of Postdoctoral Support
* Ranking of institutions is at the time of receipt of the doctorate.** Data for the trainee/fellow groups with 1–9 months and 10–12 months of Kirschstein-NRSA postdoctoral support are omitted because of the small number of observations.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.2.
ORC Macro 32
Only about a quarter of those with Kirschstein-NRSA postdoctoral support had Kirschstein-NRSA predoctoral support.
Those who applied unsuccessfully for Kirschstein-NRSA postdoctoral fellowships generally had the same level of Kirschstein-NRSA predoctoral support as Kirschstein-NRSA fellows.
Proportionally more trainees than fellows with 10–12 months of support had Kirschstein-NRSA predoctoral support.
Conversely, proportionally more fellows than trainees with 25–36 months of postdoctoral support had Kirschstein-NRSA predoctoral support.
2.4. TO WHAT EXTENT DID INDIVIDUALS RECEIVE KIRSCHSTEIN-NRSA SUPPORT DURING THEIR GRADUATE TRAINING?
The presence and extent of Kirschstein-NRSA predoctoral support may be an important factor in distinguishing among various postdoctoral populations. Kirschstein-NRSA predoctoral support occurs primarily in the form of traineeship appointments, which are made by the institutions and departments that have Ruth L. Kirschstein NRSA Institutional Research Training Grants.46 We expect that receiving such predoctoral support is more advantageous than self-support or other types of support (for example, teaching assistantships) because it allows students the freedom to focus on their research. It may also be important in that it reflects an investment in particular students by the departments administering the training grant, and as such is a recognition of the student’s potential contribution to the field. As indicated in several previously cited studies, such support may be effective in directing individuals toward postgraduate work and eventually to successful careers. In describing the study populations, we will focus on the degree to which the various study groups differ in terms of receipt of this type of support.
Figure 2.5 compares new research doctorates who received Kirschstein-NRSA predoctoral support by the type of Kirschstein-NRSA postdoctoral support they received. The data indicate that:
About one-quarter of these Kirschstein-NRSA postdoctoral recipients also received Kirschstein-NRSA predoctoral funding. This amount ranged from 23 percent for postdoctoral trainees to 29 percent for those who were supported by both traineeships and fellowships during their postdoctoral tenure.
Relative to Kirschstein-NRSA postdoctoral trainees, a slightly higher (but not notable) proportion of Kirschstein-NRSA postdoctoral fellows had Kirschstein-NRSA predoctoral support (by a 26- to 23-percent margin).
Unsuccessful Kirschstein-NRSA fellowship applicants had a level of support similar to that of successful applicants.
46 In her study of Kirschstein-NRSA predoctorates, Pion indicated that 93 percent of the individuals were supported by traineeships as opposed to fellowships. In the biological sciences, the percentage was 95 percent.
ORC Macro 33
Figure 2.5. Percentage of New Research Doctorates Receiving Kirschstein-NRSA Predoctoral Support, by Type and Receipt of Kirschstein-NRSA Postdoctoral Support*
* Very few individuals who chose not to apply for a fellowship and did not receive a traineeship received support from NRSA; they were therefore omitted from this figure.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.3.
Figure 2.6 compares postdoctoral groups by months of postdoctoral support and by type of Kirschstein-NRSA predoctoral support. The data show that whereas trainees who had 10–12 months of postdoctoral support were more likely than corresponding fellows to have had Kirschstein-NRSA predoctoral support (by a 7-percent margin), trainees who had 25–36 months of postdoctoral support were less likely than fellows to have had Kirschstein-NRSA predoctoral support (by a 10-percent margin). This pattern is notable and indicates a distinct difference between fellows and trainees in their profiles with respect to this measure.47
47 It should not be inferred from figure 2.6 that individuals with postdoctoral fellowships and NRSA predoctoral support are supported for extensive periods of time. In fact, a majority of fellows who had NRSA predoctoral support remained in their postdoctoral status for 24 months or less.
ORC Macro 34
Those with Kirschstein-NRSA postdoctoral support tended to spend less time in pursuit of the research doctorate (both in elapsed and enrolled time) than those who did not have such support.
Kirschstein-NRSA fellows tended to spend less time in pursuit of the doctorate than trainees.
Kirschstein-NRSA trainees seemed to spend more time enrolled than other groups.
Kirschstein-NRSA postdoctoral recipients with more months of postdoctoral support tended to spend more time in pursuit of the research doctorate than those with fewer months of support.
Figure 2.6. Percentage of Kirschstein-NRSA Postdoctorates Receiving Kirschstein-NRSA Predoctoral Support, by Type and Months of Kirschstein-NRSA Postdoctoral Support
* Very few individuals were supported by both postdoctoral traineeships and fellowships for less than 24 months. Because the denominators are less than five, the results would be misleading if compared with the other groups. These groups were therefore eliminated from this figure. Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.3.
2.5. HOW MUCH TIME DID INDIVIDUALS SPEND IN PURSUIT OF THE DOCTORATE?
A background variable explored in previous studies that may be indicative of potential postdoctoral career success is the length of time spent pursuing the doctoral degree. We might expect that those spending less time pursuing a degree are more focused and organized, which may be traits important for success in obtaining research funding, publishing, and conducting research. In looking at time spent in pursuit of the doctorate, elapsed time must be distinguished from enrolled time. Elapsed time is captured by examining the number of months from conferral of the bachelor’s degree to completion of the doctorate. This measure would include time not spent actively working on the degree and may reflect circumstances in which the individual sought employment prior to entering a doctoral program or during a hiatus in his
ORC Macro 35
or her studies. Enrolled time, on the other hand, represents the time that the individual is actively pursuing the doctoral degree.
2.5.1. Time to the Doctorate
Table 2.2 provides information on the elapsed time (in months) to the doctorate in terms of quartile statistics. The data indicate that:
Kirschstein-NRSA postdoctoral support recipients completed their doctorates in less time than other research doctorates. This is particularly true when we focus on the median and the third quartile statistics. In the first case, the difference is 12 months (84 months versus 96 months, respectively), and in the second case the difference is 36 months (108 months versus 144 months, respectively).
Kirschstein-NRSA fellows generally completed their doctoral degree requirements in less time than unsuccessful fellowship applicants. About 50 percent of the Kirschstein-NRSA postdoctoral fellows completed their doctorates in 72 months, whereas the same proportion of unsuccessful fellowship applicants completed their doctorates in 84 months.
The same was true for Kirschstein-NRSA fellows versus Kirschstein-NRSA trainees: About 50 percent of the Kirschstein-NRSA postdoctoral fellows completed their doctorates in 72 months, whereas the same proportion of trainees completed their doctorates in 84 months.
Table 2.2. Quartile Statistics Expressed in Elapsed Time to the Doctorate, by Type and Receipt of Kirschstein-NRSA Postdoctoral Support
Type and Receipt of Postdoctoral Support
Elapsed Time in Months
Minimum First Quartile Median Third Quartile Maximum
All Kirschstein-NRSA postdoctorates 24 72 84 108 >225
Kirschstein-NRSA fellows 24 60 72 96 >225Kirschstein-NRSA trainees 24 72 84 120 >225Kirschstein-NRSA trainee/fellows 36 72 84 96 >225Unsuccessful fellowship applicants 24 72 84 108 >225
Other research doctorates 24 72 96 144 >225
Sources: DRF, CGAF, TFF
Figure 2.7 presents another perspective on elapsed time to the doctorate. In this figure we focus on describing differences for those who spent more than 72 months (6 years) in pursuit of the doctorate. The data indicate that:
Of those individuals who did not pursue Kirschstein-NRSA funding, approximately 70 percent spent more than 72 months pursuing the doctorate. For those who received a Kirschstein-NRSA postdoctoral position, the equivalent percentage was about 57 percent.
ORC Macro 36
Proportionally fewer Kirschstein-NRSA postdoctoral fellows than unsuccessful fellowship applicants spent more than 72 months pursuing a doctorate. About 48 percent of the fellows spent more than 72 months, while about 54 percent of the unsuccessful applicants did so.
Proportionally fewer Kirschstein-NRSA fellows than Kirschstein-NRSA trainees and trainee/fellows spent more than 72 months in pursuit of the doctorate. Of these groups, 48 percent of fellows, 62 percent of trainees, and 52 percent of those with traineeships and fellowships spent more than 72 months in elapsed time.
Overall, the data indicate that Kirschstein-NRSA postdoctoral support recipients, particularly fellows, tended to complete their doctoral studies earlier than their respective comparison groups.
Figure 2.7. Percentage of New Research Doctorates With More Than 72 Months From Receipt of a Bachelor’s Degree to Receipt of the Doctorate, by Type and Receipt of
Kirschstein-NRSA Postdoctoral Support
Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.4.
Figure 2.8 presents the relationship between time to the doctorate among postdoctoral fellows and trainees by time spent in a postdoctoral training program. It is apparent that regardless of the amount of time spent in postdoctoral traineeships, relatively more postdoctoral trainees spent more than 72 months pursuing their doctorates than the other two groups. However, the data also indicate (by the upward trend as months with postdoctoral support increase) that groups who were supported by Kirschstein-NRSA for more months also seemed to take longer to complete their doctorates. For example, for those with 10–12 months of Kirschstein-NRSA postdoctoral support, the respective percentages of those taking more than 72 months to obtain their
ORC Macro 37
doctorates are 45 percent for fellows and 60 percent for trainees. For postdoctoral support recipients with 25–36 months of support, the corresponding percentages are 53 and 65 percent, respectively. When compared with the 10–12-month figures, the difference for fellows who had 25–36 months of support was 8 percentage points, and for trainees the difference was 5 percentage points. As was observed with the other analyses by months of postdoctoral support, this information indicates that the profiles of fellows and trainees differ. It also shows that postdoctorates differ by months of support.
Figure 2.8. Percentage of Kirschstein-NRSA Postdoctorates Who Spent More Than 72 Months (6 Years) in Elapsed Time To Earn Their Doctorates, by Type and Months of Kirschstein-NRSA Postdoctoral Support
* Data for the trainee/fellow groups with 1–9 months and 10–12 months of Kirschstein-NRSA postdoctoral support are omitted because of the small number of observations.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.4.
2.5.2. Enrolled Time to the Doctorate
Enrolled time reflects the actual time the graduate student was working toward his or her doctorate from the time of enrollment. In general, the same patterns are observed for enrolled time as for time to the doctorate.
As shown in table 2.3, the distribution (using quartile statistics) seemingly varied little among fellows, those with both traineeships and fellowships, and those applying unsuccessfully for fellowships. In comparing the third quartile statistics, trainees were enrolled for the longest time (120 months), and those who did not apply for or receive Kirschstein-NRSA support were enrolled for the second longest time (96 months). When compared with the results for time to the doctorate, this indicates that other research doctorates probably had hiatuses during which they were not enrolled.
ORC Macro 38
Table 2.3. Quartile Statistics Expressed in Enrolled Months to the Doctorate, by Type and Receipt of Kirschstein-NRSA Postdoctoral Support
Type and Receipt of Postdoctoral Support
Months EnrolledMinimum First Quartile Median Third Quartile Maximum
All Kirschstein-NRSA postdoctorates 12 60 72 84 >225
Kirschstein-NRSA fellows 12 60 72 84 >225Kirschstein-NRSA trainees 24 72 84 120 >225Kirschstein-NRSA trainee/fellows 24 60 72 84 168Unsuccessful fellowship applicants 12 60 72 84 >225
Other research doctorates 24 60 72 96 >225
Sources: DRF, CGAF, TFF
Figure 2.9 presents data that show the following:
Thirty-six percent of those who applied for fellowships, received fellowships, or received traineeships spent more than 72 months enrolled in a doctoral program. This contrasts with 48 percent of those who did not apply for or receive a Kirschstein-NRSA postdoctoral appointment.
Kirschstein-NRSA fellows were notably less likely to spend more than 72 months enrolled than trainees or those with both traineeships and fellowships. The differential between Kirschstein-NRSA fellows and trainees was almost 13 percentage points (28 percent versus 41 percent).
Of the fellows, about 28 percent spent more than 72 months enrolled in a doctoral program, compared with 35 percent of those who unsuccessfully applied for a Kirschstein-NRSA fellowship.
ORC Macro 39
Figure 2.9. Percentage of New Research Doctorates Who Spent More Than 72 Months in Enrolled Time From Receipt of a Bachelor’s Degree to Receipt of the Doctorate, by Type
and Receipt of Kirschstein-NRSA Postdoctoral Support
Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.5.
Figure 2.10 presents information on those enrolled for longer than 72 months by postdoctoral group and length of time spent in postdoctoral studies. These data indicate that for every postdoctoral group, a greater percentage of trainees were enrolled for more than 72 months than fellows and trainee/fellows. As noted in figure 2.8, there is a trend suggesting that as length of postdoctoral support increases, proportionally more individuals were enrolled in their doctoral programs for more than 72 months. In other words, this result suggests that individuals who take longer to complete their doctorates will receive Kirschstein-NRSA postdoctoral support for longer periods.
ORC Macro 40
Following the national trend, there has been a decline in the percentage of men who receive Kirschstein-NRSA postdoctoral support since 1962.
Proportionally fewer Kirschstein-NRSA postdoctoral support recipients were married than other research doctorates.
Proportionally fewer Kirschstein-NRSA postdoctoral support recipients had dependents when compared with other research doctorates.
Kirschstein-NRSA fellows with fewer months of support tended to be male, be married, and have more dependents. Kirschstein-NRSA trainee characteristics are similar across the groups with different months of support.
Figure 2.10. Percentage of Kirschstein-NRSA Postdoctorates Who Spent More Than 72 Months in Enrolled Time To Earn Their Doctorates,
by Type and Months of Kirschstein-NRSA Postdoctoral Support
* Data for the trainee/fellow group with 1–9 months of Kirschstein-NRSA postdoctoral support are omitted because of the small number of observations.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.5.
2.6. WHAT ARE THE DEMOGRAPHICS OF DOCTORATE RECIPIENTS?
We expect to see significant differences between the representation of men and women with regard to Kirschstein-NRSA postdoctoral funding, probably due to the relative representation of men and women in doctoral programs during the study period, which spanned almost 20 years.
In addition to gender, family obligations might be expected to play an important role in the decision to pursue opportunities other than a postdoctoral traineeship or fellowship, which may not provide sufficient income to support a family. In this section, we explore marital status and the number of dependents as indicators relating to the impact of family obligations.
As shown in table 2.4, the data indicate that about 67 percent of Kirschstein-NRSA postdoctoral recipients are male. There was little difference in the proportion of men among fellows (69 percent), trainees (64 percent), and those with both traineeships and fellowships (66 percent). There is virtually no difference in the proportion of men among those who received fellowships, those who applied for but failed to obtain a fellowship, and those who did not apply or receive NRSA support.
ORC Macro 41
Table 2.4 includes all individuals who received their doctorates in the period starting in 1965 and ending in 1992. Trend information from the DRF shows that the proportion of females in the life sciences has increased dramatically over the years.48 We might expect that the distribution of Kirschstein-NRSA postdoctoral support among males and females would also have changed. In figure 2.11, we present the proportion of new research doctorates who are male. The data indicate a slight downward trend (from more than 70 percent in 1975 to less than 60 percent in 1990).
Table 2.4. Number and Percentage of Research Doctorate Recipients, by Gender and Type of Postdoctoral Support
Type of Postdoctoral Support Male Female Total# % # % #
All Kirschstein-NRSA postdoctorates 15,796 66.6 8,162 33.4 23,709Kirschstein-NRSA fellows 5,309 68.8 2,407 31.2 7,716Kirschstein-NRSA trainees 8,938 64.4 4,950 35.6 13,888Kirschstein-NRSA trainee/fellows 1,549 65.8 805 34.2 2,354Unsuccessful fellowship applicants 6,592 68.1 3,085 31.9 9,681Other research doctorates 5,258 64.4 2,906 35.6 8,164
Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.6.
Figure 2.11. Percentage of Research Doctorates Who Are Male, 1975–1992
48 Trends in the Early Careers of Life Scientists. (1998). Commission on Life Scientists, Office of Scientific and Engineering Personnel, National Research Council, National Academy of Sciences. Washington, DC.
ORC Macro 42
Figure 2.12 provides information on the percentage of men in the three postdoctoral groups by months of postdoctoral support. The data show an interesting pattern: the percentage of men receiving fellowship support declines as months of support increase. For those with 10–12 months of support, about 73 percent of the fellows were men, while for those with 25–36 months of support, 65 percent were men. This would mean that men are, relatively speaking, in fellowships for shorter periods of time when compared with women.
Figure 2.12. Percentage of Kirschstein-NRSA Male Postdoctorates, by Type and Months of Postdoctoral Support
* Data for the trainee/fellow groups with 1–9 months and 10–12 months of Kirschstein-NRSA postdoctoral support are omitted because of the small number of observations.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.6.
In table 2.5, marital status at the time the research doctorate degree was conferred is examined. The data show that relatively more Kirschstein-NRSA fellows and trainees were unmarried than those who did not apply for or receive Kirschstein-NRSA support (46 percent to 34 percent). When successful and unsuccessful fellowship applicants are compared, the former group was slightly less likely to be married (54 percent to 56 percent). There is virtually no difference between fellows and trainees, while those who received both sources of support showed a higher likelihood of being unmarried.
Table 2.5. Number and Percentage of Research Doctorate Recipients, by Marital Status at Receipt of the Doctorate and Type and Receipt of Kirschstein-NRSA Postdoctoral Support
ORC Macro 43
Type and Receipt of Postdoctoral SupportMarried Unmarried
TotalTotal Total# % # % #
All Kirschstein-NRSA postdoctorates 12,298 53.6 10,644 46.4 22,942Kirschstein-NRSA fellows 4,015 54.1 3,402 45.9 7,417Kirschstein-NRSA trainees 7,133 53.8 6,126 46.2 13,259Kirschstein-NRSA trainee/fellows 1,150 50.8 1,116 49.3 2,266Unsuccessful fellowship applicants 5,215 56.0 4,098 44.0 9,313Other research doctorates 5,303 66.2 2,708 33.8 8,011
Sources: DRF, CGAF, and TFF
Figure 2.13 presents information on Kirschstein-NRSA postdoctoral support recipients by months of support. These data show that for all postdoctoral fellows, the relative number who are married declines as months of support increase.49 The relative difference is particularly notable when the 10–12 month group is contrasted with the 25–36 month group. The proportion for trainees remains relatively stable across the month-of-support groups, while the proportion of fellows declines.
Figure 2.13. Percentage of Kirschstein-NRSA Postdoctorates Who Are Married at Receipt of the Doctorate, by Type and Months of Kirschstein-NRSA Postdoctoral Support
* Data for the trainee/fellow groups with 1–9 months and 10–12 months of Kirschstein-NRSA postdoctoral support are omitted because of the small number of observations.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.7.
Table 2.6 presents information on individuals’ number of dependents at the time they received their doctorates. Those who did not apply for postdoctoral support (other research doctorates) were far more likely to have dependents. In all, 49 percent of those who did not apply had no dependents, compared with more than 65 percent for all Kirschstein-NRSA applicant and postdoctoral groups. There are very small differences among the three Kirschstein-NRSA postdoctoral award groups.
49 Again, marital status is measured at receipt of the doctorate. That status may change during the postdoctorate.
ORC Macro 44
Table 2.6. Number and Percentage of New Research Doctorate Recipients, by Number of Dependents at Receipt of the Doctorate and Type and Receipt of Kirschstein-NRSA Postdoctoral Support
Type and Receipt of Postdoctoral Support
Number of DependentsTotal
No Dependents One Dependent Two or More Dependents
# % # % # % #All Kirschstein-NRSA postdoctorate recipients 13,182 65.1 3,893 19.2 3,173 15.7 20,248
Kirschstein-NRSA fellows 4,317 66.7 1,255 19.4 904 14.0 6,476Kirschstein-NRSA trainees 7,483 63.6 2,268 19.3 2,011 17.1 11,762Kirschstein-NRSA trainee/fellows 1,382 68.8 370 18.4 258 12.8 2,010
Unsuccessful fellowship applicants 5,102 63.7 1,578 19.7 1,331 16.6 8,011
Other research doctorates 3,293 48.7 1,328 19.6 2,143 31.7 6,764Sources: DRF, CGAF, TFF
Figure 2.14 presents the percentage of postdoctoral support recipients with no dependents. As the data show, the percentage of fellows with no dependents increases as months of support increase. Whereas 60 percent of Kirschstein-NRSA fellows who had 10–12 months of support had no dependents, 72 percent of fellows who had 25–36 months of support had no dependents.
Figure 2.14. Percentage of Kirschstein-NRSA Postdoctorates Who Have No Dependents at Receipt of the Doctorate, by Type and Months of Kirschstein-NRSA Postdoctoral Support
* Data for the trainee/fellow groups with 1–9 months and 10–12 months of Kirschstein-NRSA postdoctoral support are omitted because of the small number of observations.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.8.
ORC Macro 45
About four-fifths of all Kirschstein-NRSA postdoctoral support recipients expressed interest in postdoctoral studies shortly after receiving their doctorates. This compares with 24 percent for other research doctorates.
Kirschstein-NRSA postdoctoral fellows were more likely than trainees to indicate that they had plans pursue postdoctoral studies (90 percent and 79 percent, respectively).
2.7. DID NEW DOCTORATE RECIPIENTS HAVE PLANS TO CONTINUE WITH POSTDOCTORAL STUDY?
Upon receiving a doctorate, individuals are asked about their postdoctoral plans through the SED. Since a formal postdoctoral appointment is critical for securing an academic research appointment in the biological sciences, we would hypothesize that most individuals receiving doctorates in the biological sciences who are interested in academic research careers will pursue a postdoctoral appointment. Individuals who do not pursue a postdoctorate may be interested in nonacademic positions or faculty positions at nonresearch institutions.
Figure 2.15 provides information about new doctorate recipients’ plans by type of Kirschstein-NRSA postdoctoral support. The data show that:
About 83 percent of those who had Kirschstein-NRSA postdoctoral support indicated that they had plans to pursue a postdoctoral position, compared with 24 percent of other research doctorates.
Ninety percent of Kirschstein-NRSA postdoctoral fellows planned to pursue postdoctoral studies, compared with 79 percent of those applying unsuccessfully for fellowships.
Kirschstein-NRSA fellows were more likely to plan to pursue postdoctoral studies than Kirschstein-NRSA trainees (90 percent and 79 percent, respectively).
It should be noted that plans and behavior were not consistent for a notable percentage of individuals. For example, as indicated above, 24 percent of those who did not apply for or receive NIH postdoctoral funding (other research doctorates) had postdoctoral plans. This percentage could reflect individuals who applied to NSF or other agencies or non-NIH organizations for postdoctoral funding as well as those who considered but did not follow up on postdoctoral plans. Conversely, 12 percent of those who had no postdoctoral plans actually applied to NIH (successfully or unsuccessfully) or were appointed to a traineeship.
ORC Macro 46
Figure 2.15. Percentage of New Research Doctorate Recipients Who Indicated Plans To Pursue Postdoctoral Study, by Type and Receipt of Kirschstein-NRSA Postdoctoral
Support
Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.9.
Figure 2.16 highlights the finding that relatively more fellows and trainee/fellows indicated that they had plans for postdoctoral study than trainees and shows that interest as expressed through their plans was highest among individuals who had postdoctoral support for a longer period. This result indicates that plans for postdoctoral study correlate with more subsequent years of Kirschstein-NRSA support. For fellows, 84 percent of those who had 10–12 months of support and 92 percent of those who had 25–36 months of support had plans for pursuing postdoctoral studies. For trainees, the corresponding percentages were 75 percent and 83 percent.
ORC Macro 47
Figure 2.16. Percentage of Kirschstein-NRSA Postdoctorates Who Had Plans To Pursue Postdoctoral Study, by Type and Months of Kirschstein-NRSA Postdoctoral Support
* Data for the trainee/fellow groups with 1–9 months and 10–12 months of Kirschstein-NRSA postdoctoral support are omitted because of the small number of observations.Sources: DRF, CGAF, TFFNote: Further details are presented in appendix table A.9.
2.8. SUMMARY AND CONCLUSIONS
2.8.1. Summary
This chapter provides a description of differences among research doctorates for factors related to the characteristics associated with their predoctoral training, personal situations at the time of receipt of the doctorate, and demographics. Within the population of research doctorates, we compared these factors by examining involvement in postdoctoral training by research doctorates and the type of training received. We have summarized our findings in our responses to the following questions (also see table 2.9 for a summary of the results from this chapter):
How are research doctorates receiving Kirschstein-NRSA postdoctoral support different from other research doctorates?
Since Kirschstein-NRSA postdoctoral funding is the predominant source of funding for postdoctoral training in the biomedical sciences, we expect this source of funds to be provided to individuals with a great potential for success in research careers. When we focus on comparing characteristics of individuals who received Kirschstein-NRSA postdoctoral funding and those who did not pursue such funding,50 we find the following:
By an 82-percent to 73-percent margin, Kirschstein-NRSA postdoctoral support recipients, compared with those without such funding, received their research doctorates from institutions that are characterized by the Carnegie Classification as the most research
50 As indicated previously, individuals who did not receive NRSA postdoctoral support may have received support from other Federal agencies or from private organizations. Some may even be receiving support from NIH research grants.
ORC Macro 48
intensive. The margin is about 9 percent in the top 100 institutions based on NIH research funding.
Kirschstein-NRSA postdoctoral support recipients generally spent less time (elapsed or enrolled) in pursuit of their research doctorates than those who received no Kirschstein-NRSA postdoctoral support.
Proportionally fewer Kirschstein-NRSA postdoctoral support recipients were married and had dependents than those who neither applied for nor received Kirschstein-NRSA postdoctoral funds. There was a 2-percent difference between these two groups in gender representation.
About 83 percent of the Kirschstein-NRSA postdoctoral fellows and trainees indicated that they had plans (just after doctorate award) to pursue postdoctoral training, compared with 28 percent of those who neither applied for nor received Kirschstein-NRSA postdoctoral support who said that they had such plans.
How are those receiving Kirschstein-NRSA fellowships different from those who applied for but did not receive a Kirschstein-NRSA fellowship?
Kirschstein-NRSA postdoctoral fellowships are competed, and the recipients are chosen through a selection process based on submitted applications. However, other factors may further distinguish between research doctorates who were awarded a fellowship and those who were not, such as the research intensiveness of the institutions from which they received their degrees. The data indicate:
Little if any difference in the research intensiveness of the institutions from which successful and unsuccessful Kirschstein-NRSA postdoctoral fellowship applicants received their research doctorates
Little difference in the extent to which successful and unsuccessful Kirschstein-NRSA postdoctoral fellowship applicants received Kirschstein-NRSA predoctoral support
Successful fellowship applicants tended to have spent less time in pursuit of the doctorate than unsuccessful applicants. This difference is still present when considering enrolled time.
No notable differences between successful and unsuccessful fellowship candidates in the proportions who were married and who had dependents
More Kirschstein-NRSA postdoctoral fellows tended to have had plans to continue with postdoctoral studies compared with unsuccessful applicants (90 percent versus 79 percent).
How are those receiving Kirschstein-NRSA postdoctoral fellowships different from Kirschstein-NRSA postdoctoral trainees?
As indicated in this chapter, Kirschstein-NRSA postdoctoral fellowships are different than traineeships in that they are competed and more tailored to the specific research interests of the
ORC Macro 49
recipient. Therefore, we might expect some differences between fellows and trainees. The data indicate:
A very small difference in the research intensity of the institutions from which fellows and trainees received their research doctorates
In terms of receipt of Kirschstein-NRSA predoctoral support, very little difference between fellows and trainees
In general, fellows took less time (elapsed or enrolled) to earn doctorates than trainees.
Little difference between fellows and trainees in the proportions who were married or had dependents. Men are marginally more prevalent among fellows than among trainees (69 percent versus 64 percent).
Fellows tended to be more directed at receipt of the doctorate, with about 90 percent indicating interest in pursuing postdoctoral studies compared with 79 percent of the trainees.
How are those Kirschstein-NRSA postdoctoral support recipients who have more months of support different from those who have fewer months of support?
Another perspective related to characterizing Kirschstein-NRSA postdoctoral fellowship and traineeship holders is the length of support received. There are a number of reasons why individuals may end their training prior to the 2-year minimum preferred by Kirschstein-NRSA. Some individuals may decide to discontinue their postdoctoral support early for reasons that include attractive research-related job offers, changes of mind about pursuit of a research career, and family responsibilities. The data indicate:
No notable differences between Kirschstein-NRSA fellows and trainees who have 10–12 months of support, but fellows who have 25–36 months of support tend, more so than trainees, to have completed doctoral studies in research-intensive institutions.
Proportionally more trainees than fellows with 10–12 months of postdoctoral support had Kirschstein-NRSA predoctoral support. The reverse is observed for those with 25–36 months of support.
The elapsed and enrolled time to the doctorate increases as months of postdoctoral support increase.
For Kirschstein-NRSA postdoctoral fellows relative to trainees, men tended to be less represented among postdoctoral groups with more months of postdoctoral support. Again for fellows when compared with trainees, married individuals tended to be less represented than unmarried individuals in groups with more months of postdoctoral funding. Fellows with a greater number of months of postdoctoral funding also tended to have had fewer dependents than those with fewer months of funding.
ORC Macro 50
An apparent positive relationship between months of postdoctoral support and indicated plans to pursue a postdoctoral position.
ORC Macro 51
Table 2.9. Summary of Background Differences Among Research Doctorates
MeasureAll Kirschstein-
NRSA Postdoctorates
Kirschstein-NRSA
Fellows
Kirschstein-NRSA
Trainees
Kirschstein-NRSA
Trainee/Fellows
Unsuccessful Fellowship Applicants
Other Research
Doctorates
Kirschstein-NRSA Postdoctorates
Compared With Other Research Doctorates
Other Difference
s
Received doctorate from research-intensive institution (Carnegie Classification)
82% 85% 81% 81% 83% 73% Kirschstein-NRSA postdoctorates receive their doctorates from more research-intensive institutions.
None
Received doctorate from research-intensive institution (NIH funding)
67% 69% 66% 67% 66% 58% Kirschstein-NRSA postdoctorates receive their doctorates from more research-intensive institutions.
None
Received Kirschstein-NRSA predoctoral support
24% 26% 23% 29% 22% * No comparison None
Elapsed time spent in pursuit of the doctorate greater than 72 months
57% 48% 62% 52% 54% 70% Kirschstein-NRSA postdoctorates spend less elapsed time completing their doctorates.
Fellows spend less time than trainees
Enrolled time spent in pursuit of the doctorate greater than 72 months
36% 28% 41% 33% 36% 48% Kirschstein-NRSA postdoctorates spend less elapsed time completing their doctorates.
Fellows spend less time than trainees
Gender (percentage male)
67% 69% 64% 66% 68% 64% No difference None
Marital status (percentage married upon receipt of the doctorate)
54% 54% 54% 51% 56% 66% Kirschstein-NRSA postdoctorates are less likely to be married.
None
Dependency (percentage with no dependents)
65% 67% 64% 69% 64% 49% Kirschstein-NRSA postdoctorates are less likely to have dependents.
None
Plans to pursue postdoctoral studies
83% 90% 79% 91% 79% 28% Kirschstein-NRSA postdoctorates are more likely to have plans to pursue postdoctoral studies.
Fellows more likely to have plans than trainees
ORC Macro 52
* Could not be estimated from available data.
ORC Macro 53
2.8.2. Conclusions
The largest difference in predoctoral experience and background characteristics among the comparison groups, as expected, is between those who indicated that they had plans to pursue Kirschstein-NRSA postdoctoral support and those who did not.
There are several differences between those receiving Kirschstein-NRSA postdoctoral support and those neither applying for a Kirschstein-NRSA postdoctoral fellowship nor receiving traineeship support. Using these individuals as comparisons without further refinement, however, may limit their ability to provide a satisfactory test of Kirschstein-NRSA effects.
Successful fellowship applicants are distinguished from unsuccessful applicants in a number of areas; however, overall the differences between the two groups are marginal.
Differences between fellows and trainees are also marginal, with the important factors relating to interest in pursuing a postdoctorate and success in obtaining the doctorate in a timely fashion. These differences are amplified when length of time receiving Kirschstein-NRSA postdoctoral support is taken into account. Distinct differences were found between those with fewer months of postdoctoral support and those with more months of postdoctoral support. These differences suggest that Kirschstein-NRSA trainee and fellow populations have different profiles resulting in part from the length of support that they receive.
Overall the evidence seems to indicate that characteristics associated with predoctoral training differentiate Kirschstein-NRSA fellows and trainees. Most factors, such as the research intensiveness of the degree-granting institution, family responsibilities, and success in completing the degree in a timely manner, seem to be moderating influences that have an effect on both the receipt of Kirschstein-NRSA support and the length of Kirschstein-NRSA support.
ORC Macro 54
3. THE NATURE OF KIRSCHSTEIN-NRSA SUPPORT
Further research experience in a postdoctoral position is the next career step for a large number of individuals receiving research doctorates in the biological science, behavioral science, and clinical research fields. The postdoctoral experience exposes individuals to further mentoring, the possibility of participating in cutting-edge research, and an opportunity to change their study emphasis. For those with M.D.’s, it offers a chance to develop or enhance research skills. Although some individuals pursue postdoctoral opportunities offered by agencies or organizations other than NIH, Kirschstein-NRSA fellowships and traineeships play a critical role in supporting postdoctoral studies.
To explore the assumption that the Kirschstein-NRSA postdoctoral experience has an effect on career outcomes, this chapter provides a basis for understanding the nature of Kirschstein-NRSA support. The following questions will be addressed:
What type of Kirschstein-NRSA support was provided for the first and subsequent postdoctoral positions? As described in chapter 2, the Kirschstein-NRSA program may support particular individuals through one or a series of fellowships, traineeships, or both traineeships and fellowships. In general, fellowships can be awarded for a period ranging from 1 to 3 years; 1-year fellowships must be recompeted in order for an individual to continue receiving support, and multiyear fellowships require noncompeting applications every year. Traineeships are generally year-to-year appointments administered through an institutional grant. The kinds of Kirschstein-NRSA support received and the sequencing of support during postdoctoral studies may have an effect on the career outcomes of individuals. For example, we could examine the comparative career outcomes of fellows and trainees or those of individuals who cycle through a number of awards versus individuals who receive a 3-year fellowship.
How long did Kirschstein-NRSA postdoctorates receive Kirschstein-NRSA support? This question reflects further on the support patterns of individuals receiving Kirschstein-NRSA support. There is an explicit desire on the part of NIH that individuals remain committed to a postdoctoral arrangement for 2 years. This commitment is seen as fostering the skills necessary for advancing the individual’s career and making an additional commitment to a research career.51 It can be hypothesized that the longer the appointment (within the mandated limits), the more the individual will be directed toward research-oriented outcomes.
51 “Studies have shown that the length of the appointment to a training grant for postdoctoral trainees with health-professional degrees strongly correlates with subsequent application for and receipt of independent NIH research support. Training grant program directors, therefore, should limit appointments to individuals who are committed to a career in research and plan to remain on the training grant or in a non-Kirschstein-NRSA research experience for a cumulative minimum of 2 years. It has also been shown that transition to independent support is related to career success. Therefore, program directors should encourage postdoctoral trainees to apply for the individual postdoctoral fellowships (F32) or mentored career development awards (K awards).” NIH National Research Service Award Institutional Research Training Grants (T32), Release Date: May 16, 2002. Expiration Date: December 31, 2005, unless reissued. PA NUMBER: PA-02-109.
ORC Macro 55
In which fields of research were Kirschstein-NRSA postdoctoral recipients involved? Kirschstein-NRSA fellowship applicants are specifically assessed on their proposed research, and trainee appointees become involved in specific research projects supported by the institutions/departments that are awarded training grants. This research could be a continuation of or be different from research pursued as a predoctoral student. In any case, the area of research will probably have an effect on subsequent career outcomes and may be a differentiator in terms of outcomes. Therefore, the areas of research, even if defined in very broad disciplinary terms, may provide us with information on the nature of Kirschstein-NRSA postdoctoral research.
What was the research intensity of the institutions in which Kirschstein-NRSA postdoctoral recipients conducted research? We expect that working in environments with accomplished researchers on cutting-edge research is more likely to result in successful research careers. The characteristics of institutions sponsoring Kirschstein-NRSA postdoctorates are expected to be important in fostering research-oriented career outcomes among postdoctorates. As was done in chapter 2, we will again use research intensity to characterize institutions.
These questions will be addressed through comparisons made among Kirschstein-NRSA postdoctoral recipients by type of degree. We will compare individuals with research doctorates with individuals with M.D.’s and those with M.D./research doctorates.52 We expect these groups to differ regarding the type and length of Kirschstein-NRSA support because motivation and the extent of research skills and training will vary among the groups.
The information contained in this chapter derives from the CGAF and TFF for all Kirschstein-NRSA recipients studied in chapter 2. That population includes 23,969 research doctorates who received Kirschstein-NRSA postdoctoral fellowship or traineeship support between FY 1975 and FY 1992. In the initial analyses in section 3.1, we also include 9,725 research doctorates who were unsuccessful fellowship applicants. In addition, we will examine 7,712 M.D.’s who have received Kirschstein-NRSA postdoctoral fellowship or training support or have applied for Kirschstein-NRSA postdoctoral fellowship support.
3.1. WHAT KIND OF KIRSCHSTEIN-NRSA SUPPORT WAS RECEIVED FOR THE FIRST AND SUBSEQUENT POSTDOCTORAL POSITIONS?
Our expectation is that a large proportion of individuals receiving their research doctorates would seek a postdoctorate within 1–2 years of receipt of their degrees. This expectation is based on the view that postdoctoral training is a continuation of predoctoral training and the next step in pursuing a research career. However, this expectation may be different for individuals with professional degrees who have developed an interest in pursuing research rather than clinical work. For these individuals, we may expect a greater timespan between receipt of the doctorate and postdoctoral training because M.D.’s and other health professionals may be involved in internships or residencies prior to switching to a research focus.
52 Research doctorates refer to individuals who have earned Ph.D.’s or any equivalent research-oriented degrees. Henceforth, we refer to all research doctorates in this group as Ph.D.’s.
ORC Macro 56
Figure 3.1 shows the timespan from receipt of the research doctorate to the first Kirschstein-NRSA postdoctoral fellowship or traineeship.53 The data indicate that most Kirschstein-NRSA postdoctoral support recipients obtain support within the first year after receipt of their research doctorates (65 percent). However, a substantial number of individuals obtain support 2 or more years after receiving their doctorates (29 percent), and some first receive Kirschstein-NRSA postdoctoral support 6 or more years after receiving their doctorates (5 percent). When type of degree is considered, a relatively high percentage of those with dual professional and research doctorates receive support prior to the receipt of their research doctorates (29 percent).
Figure 3.1. Percentage of Kirschstein-NRSA Postdoctorate Recipients by Elapsed Time (in Years) Between Award of the Research Doctorate and First Kirschstein-NRSA
Postdoctoral Support, by Type of Degree
Sources: CGAF, TFF
Unlike Kirschstein-NRSA postdoctoral traineeship appointments, Kirschstein-NRSA fellowships offer opportunities to postdoctoral students to pursue their own research interests under the sponsorship of a noted researcher. Table 3.1 shows the distribution of Kirschstein-NRSA postdoctoral recipients by the type of Kirschstein-NRSA support received. This table only considers the postdoctoral mechanisms examined in this study (F32, F35, and T32). Most postdoctoral support recipients were Kirschstein-NRSA trainees (64 percent), and about one-quarter were Kirschstein-NRSA postdoctoral fellows with F32 support (27 percent). Those who held both fellowships and traineeships at different times accounted for 9 percent of the total.
Of those with research doctorates, 33 percent received F32 support. This percentage exceeds the percentage of M.D.’s who had F32 support (9 percent) and the percentage of those with dual degrees who had F32 fellowships (21 percent).54 M.D.’s were proportionately more evident among trainees than research doctorates and those with dual degrees. M.D./Ph.D.’s were
53 Individuals with only a professional doctorate (e.g., M.D.’s) were not included in this figure due to a lack of information on when they received their professional doctorates.54 It should be noted that M.D.’s in general may undertake additional training, such as internships and residencies, after receipt of their degrees. Such training may lead to very different paths for M.D.’s who become interested in research.
ORC Macro 57
marginally more highly represented among those who held both a fellowship and traineeship. The data indicate an association between certain degrees and the type of support received.
Table 3.1. Distribution of Kirschstein-NRSA Postdoctoral Support Recipients Across Award Mechanisms, by Type of Degree, FY 1975–1992
Type of Degree
Fellows
Trainees (T32)
FellowsWho AlsoReceived
TraineeshipSupport
TotalF32 F35
# % # % # % # % # %
All Kirschstein-NRSA postdoctorates
8,340 27.0 46 0.1 19,626 63.6 2,848 9.2 30,860 100.0
Research doctorates 7,201 33.4 4 0.0 12,262 56.8 2,105 9.8 21,572 100.0
M.D.’s 634 9.2 32 0.5 5,732 83.2 493 7.2 6,891 100.0
M.D./Ph.D.’s 505 21.1 10 0.4 1,632 68.1 250 10.4 2,397 100.0
Sources: CGAF, TFF
Table 3.1 provides an overall summary of the kinds of support provided to Kirschstein-NRSA postdoctoral support recipients from FY 1975 to FY 1992. These categories of support, however, consider the Kirschstein-NRSA postdoctoral experience as a unitary event or intervention. As the data show, about 9 percent of Kirschstein-NRSA postdoctoral recipients receive both traineeship and fellowship support. Within this group, our assumption was that individuals generally cycle from traineeships to fellowships and not the reverse.55 Furthermore, it is not clear whether individuals who are unsuccessful at obtaining a fellowship when they first apply subsequently obtain another type of Kirschstein-NRSA support. Thus, it would seem that the types of Kirschstein-NRSA postdoctoral support may be more nuanced than those defined by the general program categories.
To investigate these distinctions, we focus on charting the postdoctoral events of individuals interested in Kirschstein-NRSA support, as indicated by their application for a Kirschstein-NRSA fellowship or by their appointment to a traineeship. The first award or appointment after receipt of a doctoral degree is critical because it greatly influences subsequent behavior. For example, individuals who receive initial fellowship support for 3 years are assured of continuing support, while individuals who receive a 1-year fellowship may have to compete for further fellowship support. Trainees, on the other hand, are generally appointed for 1-year terms. The first award or appointment may therefore dictate the subsequent path taken by an individual.To examine these patterns, we focus on the fellowship application and traineeship appointment process. Figure 3.2 provides a flow diagram of Kirschstein-NRSA support patterns during two
55 The NIH program announcement for the T32 traineeship program promotes this idea by stating that “Training grant program directors, therefore, should limit appointments to individuals who are committed to a career in research and plan to remain on the training grant or in a non-Kirschstein-NRSA research experience for a cumulative minimum of 2 years. It has also been shown that transition to independent support is related to career success. Therefore, program directors should encourage postdoctoral trainees to apply for the individual postdoctoral fellowships (F32) or mentored career development awards (K awards).”
ORC Macro 58
award segments, which correspond to a period or award segment (usually a funding period) in which an action occurs. This figure provides a “map” for examining the data tables that follow.
The first segment occurs when an individual first applies for a fellowship or receives a traineeship appointment. Figure 3.2 defines several events for this segment:
Successful award of a Kirschstein-NRSA fellowship
1-year award Multiyear award
Receipt of a Kirschstein-NRSA traineeship appointment Unsuccessful Kirschstein-NRSA fellowship application
Individuals have the option of applying for Kirschstein-NRSA fellowship support during a second segment. In the case of multiyear fellowship awardees, individuals would be expected to submit a noncompeting continuation application.56
For the other groups, individuals would need to submit a competing fellowship application in order to continue receiving Kirschstein-NRSA postdoctoral fellowship support. Trainees and unsuccessful Kirschstein-NRSA fellowship applicants can apply for a fellowship during any of the three application periods that occur in each fiscal year.57 Previously unsuccessful applicants must present an application that addresses the weaknesses cited by the review panel in the first review. Although we can stipulate further award segments, we believe that the representation of two award segments is adequate for providing evidence of the variations inherent in Kirschstein-NRSA postdoctoral support.
56 The noncompeting continuation application is formally referred to as the 416-9, whereas the competing application is the 419-1.57 Individuals can submit Kirschstein-NRSA fellowship applications in December, April, and August of each fiscal year. The application stipulates that individuals cannot submit a reapplication while a former application is pending, nor can individuals submit an subsequent application that does not address comments made during the review of a failed previous submission.
ORC Macro 59
What are the characteristics of postdoctoral recipients during their initial Kirschstein-NRSA funding period?
Individuals initially appointed to traineeships outnumber fellows by a 3-to-1 margin.
Research doctorates are more likely to have received a fellowship position than those with professional degrees.
More than one-third of those applying for fellowships are successful (37 percent). Those with dual degrees have a much higher award rate than other applicants (57 percent).
Figure 3.2. Overview of the Kirschstein-NRSA Fellowship Application and Award and Traineeship Appointment Segments for Research Doctorates
Because Kirschstein-NRSA postdoctoral fellowships are competed, they allow individuals more freedom in pursuing research interests. From FY 1975 to FY 1992, there were 9,320 individuals who were awarded a Kirschstein-NRSA fellowship. As shown in figure 3.3, about 87 percent of first-time Kirschstein-NRSA fellowship applicants have research doctorates, 9 percent have M.D.’s, and 4 percent have dual degrees.
ORC Macro 60
Figure 3.3. Distribution of First-Time Kirschstein-NRSA Fellowship Applicants, by Degree
Sources: CGAF, TFF
The overall award rate for the first Kirschstein-NRSA fellowship application during the study period was about 37 percent. This number excludes individuals who were first appointed to a traineeship and then applied for a fellowship; this rate therefore reflects those with no previous interaction with the Kirschstein-NRSA postdoctoral program. The percentage varies from about 36 percent for research doctorates and 37 percent for those with M.D.’s to 57 percent for those with dual degrees (see table 3.2). Despite the fact that those with dual degrees constitute a very small percentage of all applicants, the award rate indicates that this population was more likely to be awarded fellowships than those with either M.D.’s or research doctorates.
Table 3.2. Number of Applicants, Awardees, and Award Rates of New Kirschstein-NRSA Postdoctoral Fellowship Applicants, by Type of Degree
Type ofDegree
New Kirschstein-NRSA Postdoctoral Fellowship Applicants
Total Applicants Received Award Award Rate
Total 21,125 7,718 36.5
Research doctorates 18,349 6,586 35.9
M.D.’s 1,831 651 36.6
M.D./Ph.D.’s 845 481 56.9
Note: Data refer to the first fellowship application. Individuals had no previous support from the Kirschstein-NRSA program.Sources: CGAF, TFF
For the initial segment, there were about 2.6 Kirschstein-NRSA postdoctoral trainees for every Kirschstein-NRSA fellow. For research doctorates, there were 1.9 trainees for every fellow, while for the M.D. and M.D./Ph.D. groups, the trainee to fellow ratio was 2.2 to 1 and 3.5 to 1,
ORC Macro 61
respectively. These figures indicate that research doctorates are less likely to be represented among those with traineeships than M.D.’s and those with dual degrees (see table 3.3).
Table 3.3. Number of Kirschstein-NRSA Postdoctorate Recipients, by Initial Supporting Mechanism and Type of Degree
Type ofDegree
Received Kirschstein-NRSA
Postdoctoral Fellowship
Appointed to a Kirschstein-NRSA
Postdoctoral Traineeship
Ratio of Trainees to Fellows
All Kirschstein-NRSA postdoctorates 7,718 19,842 2.6
Research doctorates 6,586 12,231 1.9
M.D.’s 2,651 5,926 2.2
M.D./Ph.D.’s 481 1,685 3.5
Sources: CGAF, TFF
One of the important factors expected to affect the path of individuals who succeed in obtaining fellowship funding is the length of initial support. Individuals who are funded for less than 3 years would need to apply for support and compete against a new set of applicants in order to continue their postdoctoral studies as fellows. Figure 3.4 shows the distribution of Kirschstein-NRSA postdoctoral fellows by months of support from the first competing fellowship award. The data show that about one-quarter (26 percent) of the awards are for 1 year or less, and two-fifths (42 percent) are for a period of 13–24 months. Thus, about 68 percent of all competing fellowship awards are for periods of 2 years or less. These data also show that individuals with M.D.’s tend to have fellowships that are shorter in length: About 37 percent of those with M.D.’s had Kirschstein-NRSA fellowships that were 12 months or less in length, compared with about 25 percent for the other two degree groups.
ORC Macro 62
Figure 3.4. Length of Initial Kirschstein-NRSA Competing Fellowship, by Type of Degree
Sources: CGAF, TFFNote: Further details are presented in appendix table B.1.
In the following subsections we first look at individuals who initially applied for and received a fellowship, then individuals who initially received a traineeship (who did not apply for a fellowship prior to receiving the traineeship), and finally individuals who did not succeed at their first attempt to obtain a fellowship.
3.1.1. Subsequent Kirschstein-NRSA Support Patterns for Individuals Who Initially Applied for and Received a Fellowship
Individuals can be supported under the same fellowship for up to 3 years. Individuals who receive less than 3 years of support and want to continue their postdoctoral career with a Kirschstein-NRSA fellowship must compete for a second segment of support. Table 3.4 shows the distribution of Kirschstein-NRSA fellows who were successful in their first attempt to obtain fellowship support by the subsequent activities pursued after the initial segment of fellowship support. It does not include individuals who were first appointed to traineeships (that group is considered in section 3.1.2). Table 3.4 shows that:
ORC Macro 63
What happens to individuals who are awarded a fellowship as their first source of Kirschstein-NRSA support?
About 20 percent reapply for a second fellowship, with an award rate of 72 percent.
About 5 percent are appointed to a traineeship. About 76 percent receive no further support from
Kirschstein-NRSA postdoctoral mechanisms. These individuals could have 1 or more years of support through a noncompeting mechanism.
About 20 percent of all fellows reapplied for a Kirschstein-NRSA fellowship after their initial award segment. Research doctorates were more likely than M.D.’s to apply for another fellowship (21 and 14 percent, respectively).
Approximately 5 percent of all Kirschstein-NRSA postdoctoral fellows received a Kirschstein-NRSA postdoctoral traineeship after the first segment of fellowship support.
Three-quarters of all fellows received no further Kirschstein-NRSA support after their first competing award.
These findings must be interpreted in the context of the definition of segment used above. Award segments reflect an event history and not a timeline history. That is, individuals in their first award segment may be awarded a 1-, 2-, or 3-year award. Individuals receiving no further support may be individuals with 1, 2, or 3 years of fellowship support.
Table 3.4. Number and Percentage of Kirschstein-NRSA Postdoctoral Fellows Seeking and Receiving Kirschstein-NRSA Support After Their Initial Segment of Support, by Type of Degree
Type of Degree
Total Individuals
With a Kirschstein-
NRSA Postdoctoral Fellowship
During Initial Segment
Subsequent Activity for Fellows in Second Segment
Reapplied for Fellowship
Appointed to Traineeship
No Further Kirschstein-NRSA
Support
# % # % # %
Total 7,718 1,516 19.6 367 4.8 5,835 75.6
Research doctorates 6,586 1,352 20.5 297 4.5 4,937 75.0
M.D.’s 651 92 14.1 39 6.0 520 79.9
M.D./Ph.D.’s 481 72 15.0 31 6.4 378 78.6
Sources: CGAF, TFF
As shown in table 3.4, there were 1,516 Kirschstein-NRSA postdoctoral fellows who applied for a second segment of fellowship support. Table 3.5 shows the success of these individuals by degree. Overall, about 72 percent of fellows were awarded a second fellowship. Research doctorates were the most successful (73 percent), and those with dual degrees were the least successful (60 percent). Although not explored in this analysis, an interesting question is why individuals who were successful in their first application failed to obtain fellowship funding for their second application.
ORC Macro 64
Table 3.5. Award Rate of Kirschstein-NRSA Fellows Who Recompeted for Kirschstein-NRSA Fellowship Funding, by Type of Degree
Type of Degree
Number of Kirschstein-NRSA Postdoctoral
Fellows Who Applied for a Second Segment of Fellowship Support
Awarded Kirschstein-NRSA Fellowship
Number Awarded Award Rate
Total 1,516 1,094 72.2
Research doctorates 1,352 985 72.9
M.D.’s 92 66 71.7
M.D./Ph.D.’s 72 43 59.7
Sources: CGAF, TFF
3.1.2. Subsequent Kirschstein-NRSA Support Patterns for Individuals Who Initially Received Traineeships
The majority of Kirschstein-NRSA postdoctoral support recipients are trainees. Traineeships are awarded to institutions through competing grants, and once the award is received, the institution selects individual students to receive traineeships. Almost 20,000 individuals in our study population received a Kirschstein-NRSA postdoctoral traineeship as their first support mechanism. Once a traineeship was received, individuals could apply for a fellowship at any point, receive an additional traineeship appointment, or leave their traineeship appointment during or after a period of support.
Table 3.6 shows that for individuals who were first appointed to a traineeship, 14 percent subsequently applied for a fellowship, 60 percent were reappointed to a traineeship, and 26 percent did not seek or receive further Kirschstein-NRSA support. Unlike fellows, who can receive an award for periods of up to 3 years, trainees are reappointed each year under the institutional grants. Individuals with research doctorates were more likely to subsequently apply for fellowships than M.D.’s (17 and 9 percent, respectively) and to be reappointed to a traineeship than M.D.’s (62 and 55 percent, respectively). M.D.’s were more likely to receive no further Kirschstein-NRSA support than the other degree groups. Because trainees are appointed each year, this means that that those with no further Kirschstein-NRSA support generally leave their postdoctoral appointments or receive other types of funding after 1 year of support.
ORC Macro 65
What happens to individuals who are appointed to a traineeship as their first support mechanism?
About 14 percent apply for a fellowship, with an award rate of 41 percent. Research doctorates show a higher fellowship application rate and lower award rate than other degree groups.
About 62 percent are appointed to another traineeship. Research doctorates show a higher reappointment rate than M.D.’s.
About 26 percent receive no further support from the Kirschstein-NRSA program.
Table 3.6. Number and Percentage of Kirschstein-NRSA Postdoctoral Trainees Seeking and Receiving Kirschstein-NRSA Postdoctoral Support After Their Initial Segment of Support, by Type of Degree
Type of Degree
Total Individuals With Kirschstein-
NRSA Postdoctoral Traineeship
Support During Initial Segment
Subsequent Activity for Trainees in Second Segment
Applied for Fellowship
Reappointed to Traineeship
No Further Kirschstein-NRSA
Support
# % # % # %
Total 19,842 2,725 13.7 11,887 59.5 5,230 26.4
Research doctorates 12,231 2,021 16.5 7,592 62.1 2,618 21.4
M.D.’s 5,926 501 8.5 3,266 55.1 2,159 36.4
M.D./Ph.D.’s 1,685 203 12.1 1,029 61.1 453 26.9
Sources: CGAF, TFF
Of the 2,725 trainees who applied for a Kirschstein-NRSA fellowship after receiving traineeship support, 1,105 individuals (about 41 percent) were awarded one (see table 3.7). This award rate is about 32 percent lower than the success rate for Kirschstein-NRSA fellows who subsequently applied for a second fellowship (see table 3.5). Award rate by degree varies, with about 38 percent of research doctorates receiving a fellowship, compared with 47 percent for M.D.’s and 48 percent for those with dual degrees.
Table 3.7. Number of Applicants, Awardees, and Award Rates for Trainees Who Competed for Kirschstein-NRSA Fellowship Funding After Initial Traineeship Appointment, by Type of Degree
Type of Degree
Number of Kirschstein-NRSA Postdoctoral Trainees Who
Applied for Fellowship Support After Initial
Traineeship Appointment
Awarded Kirschstein-NRSAFellowship
Number Awarded Award Rate
Total 2,725 1,105 40.5
Research doctorates 2,021 775 38.3
M.D.’s 501 233 46.5
M.D./Ph.D.’s 203 97 47.8
Sources: CGAF, TFF
ORC Macro 66
What happens to individuals who are unsuccessful in their first attempt to obtain Kirschstein-NRSA postdoctoral fellowship support?
About 27 percent reapply, with an award rate of 27 percent. Those with dual degrees show a higher fellowship application rate and higher award rate than other degree groups.
About 12 percent are appointed to a traineeship. About 60 percent do not apply for or receive
further Kirschstein-NRSA support.
3.1.3. Subsequent Kirschstein-NRSA Support Patterns for Individuals Who Failed To Obtain Fellowships on Their First Attempt
From the information provided in table 3.2, we can determine that 63 percent of those individuals applying for a Kirschstein-NRSA fellowship fail on their first attempt. Some reapply, some obtain a traineeship appointment instead, and others may obtain non-Kirschstein-NRSA support or end their attempt to obtain Kirschstein-NRSA postdoctoral support. Table 3.8 provides information on what happens to these individuals after the initial failed attempt. Overall, about 27 percent of individuals who were initially unsuccessful in obtaining a fellowship reapplied. About 12 percent of unsuccessful fellowship applicants were successful in obtaining a traineeship appointment. Nearly 60 percent did not receive further Kirschstein-NRSA postdoctoral support. In other words, they applied once, never applied again, and never received traineeship support.
When these initially unsuccessful fellowship applicants are compared by degree, those with dual degrees (M.D./Ph.D.’s) reapplied for fellowships or were appointed to traineeships at a higher rate than those with only M.D.’s or research doctorates. About 36 percent of those with dual degrees reapplied for fellowships, compared with 28 percent of individuals with research doctorates and 21 percent of M.D.’s (see table 3.8).
Table 3.8. Number of Unsuccessful First-Time Fellowship Applicants Seeking and Receiving Kirschstein-NRSA Postdoctoral Support, by Type of Degree
Type of Degree
Total Unsuccessful
Fellowship Applicants
During Initial Segment of
Support
Subsequent Activity for Unsuccessful Fellowship Applicants
Reapplied for Fellowship
Appointed to Traineeship
No Subsequent Kirschstein-NRSA
Support
# % # % # %
Total 13,504 3,688 27.3 1,655 12.3 8,161 60.4
Research doctorates 11,763 3,280 27.9 1,360 11.6 7,123 60.6
M.D.’s 1,180 243 20.6 183 15.5 754 63.9
M.D./Ph.D.’s 364 131 36.0 112 30.8 121 33.2
Sources: CGAF, TFF
ORC Macro 67
Of the individuals who were unsuccessful in their first attempt to obtain Kirschstein-NRSA fellowship support, 3,688 applied a second time. Of these individuals, 27 percent were successful in obtaining a fellowship (see table 3.9). Those with dual degrees were the most successful, with a success rate of 44 percent. M.D.’s were the least successful, with about 20 percent obtaining a fellowship on their second attempt.
Table 3.9. Award Rates of Unsuccessful Kirschstein-NRSA Fellowship Applicants Who Subsequently Competed for Kirschstein-NRSA Fellowship Funding, by Type of Degree
Type ofDegree
Number of Unsuccessful Kirschstein-NRSA
Postdoctoral Fellowship Applicants Who Applied
Again for Fellowship Support
Awarded Kirschstein-NRSAFellowship
Number Awarded Award Rate
Total 3,688 980 26.6
Research doctorates 3,280 874 26.6
M.D.’s 243 49 20.2
M.D./Ph.D.’s 131 57 43.5
Sources: CGAF, TFF
This section described how individuals—particularly Kirschstein-NRSA fellows—are funded through the first two award segments of funding. Interesting comparisons for further analysis may include:
Individuals who receive fellowship funding and then apply unsuccessfully for a second segment of funding versus those who are successful in obtaining further support
Individuals who apply unsuccessfully and receive no further support versus individuals who apply unsuccessfully but subsequently are successful in obtaining a fellowship
Individuals who obtain fellowship support through several sequential competing awards versus those who obtain more than 1 year of fellowship support through a single competition.
These comparisons speak to the need to understand how individuals interact with the Kirschstein-NRSA training program over their entire postdoctoral study period.
ORC Macro 68
3.2. WHAT WAS THE LENGTH OF KIRSCHSTEIN-NRSA POSTDOCTORAL SUPPORT?
The length of Kirschstein-NRSA support has been reported to be associated with higher research productivity,58 and there are disincentives (in the form of service paybacks) for individuals who leave within the first 2 years of their postdoctorate,59 particularly if they leave for nonresearch-related employment. For most individuals interested in pursuing a research career at a university or similar research institution, their length of support may reflect job market conditions and the types of job opportunities available, an interest in continued involvement in a particular type of research as a postdoctorate, or a perceived need to master certain skills.
In general, the median period for which Kirschstein-NRSA postdoctorates received support is 24 months (see table 3.10). This figure does not vary substantially by type of degree; however, the median for those who received both a traineeship and fellowship is 36 months, approximately 1 year longer than those who received either a fellowship or traineeship. It should be noted that these figures represent the entire history of support, not just a single award or appointment.
58 NIH National Research Service Award Institutional Research Training Grants (T32), Release Date: May 16, 2002. Expiration Date: December 31, 2005, unless reissued. PA Number: PA-02-109.59 For every month in the second year of Kirschstein-NRSA postdoctoral support, an individual reduces the repayment obligation incurred in the first year by a month. At the end of 2 years, the obligation is repaid. There are exceptions for individuals who pursue research or teaching careers.
ORC Macro 69
What was the length of Kirschstein-NRSA postdoctoral support?
About 50 percent of Kirschstein-NRSA fellows and trainees had 2 years of support; for those with both traineeships and fellowships, the median period of support was about 3 years.
Individuals who were supported by both traineeships and fellowships during their postdoctoral studies were supported for a longer period of time.
Table 3.10. Median Number of Months of Kirschstein-NRSA Postdoctoral Support, by Type of Support and Degree
Type of Support
Type of DegreeTotal
Ph.D. M.D. M.D./Research Doctorate Other
MedianMonths # Median
Months # MedianMonths # Median
Months # MedianMonths #
All Kirschstein-NRSA postdoctorates
24 21,572 24 6,891 24 2,397 24 1,597 24 32,457
Kirschstein-NRSA fellows 24 7,205 24 666 24 515 24 340 24 8,726
Kirschstein-NRSA trainees 24 12,262 22 5,732 24 1,632 21 1,171 24 20,797
Kirschstein-NRSA trainee/fellows
36 2,105 36 493 36 250 36 86 36 2,934
Sources: CGAF, TFF
Another perspective on the length of Kirschstein-NRSA support is provided by figures 3.5a–3.5c, which show the distribution of the Kirschstein-NRSA postdoctoral population by months of Kirschstein-NRSA support. Figure 3.5a, which focuses on research doctorates, shows different patterns for fellows, trainees, and individuals who held both traineeships and fellowships during their postdoctoral tenure. In particular, there are proportionally more research fellows in the 13–24-month and 25–36-month Kirschstein-NRSA support groups than there are trainees. Proportionally, more individuals with both traineeships and fellowships received more than 24 months of support than did fellows and trainees; about 30 percent of this group received support beyond 36 months. This pattern is replicated for individuals with M.D.’s and those with dual degrees (see figures 3.5b and 3.5c).
ORC Macro 70
Figure 3.5a. Percentage of Research Doctorate Kirschstein-NRSA Postdoctoral Support Recipients, by Months and Type of Kirschstein-NRSA Support
Sources: CGAF, TFFNote: Further details are presented in appendix table B.2.
Figure 3.5b. Percentage of M.D. Kirschstein-NRSA Postdoctoral Support Recipients,by Months and Type of Kirschstein-NRSA Support
Sources: CGAF, TFFNote: Further details are presented in appendix table B.2.
ORC Macro 71
Figure 3.5c. Percentage of Dual Degree Kirschstein-NRSA Postdoctoral Support Recipients, by Months and Type of Kirschstein-NRSA Postdoctoral Support
Sources: CGAF, TFFNote: Further details are presented in appendix table B.2.
3.3. IN WHAT FIELDS WERE RECIPIENTS DOING RESEARCH WHILE RECEIVING KIRSCHSTEIN-NRSA SUPPORT?
The kinds of research that individuals pursued while receiving Kirschstein-NRSA postdoctoral support include the following broad research categories:
Biomedical sciences Chemistry Clinical sciences Behavioral and social sciences Other disciplines
With regard to all Kirschstein-NRSA postdoctoral recipients, the data show that research doctorates are most likely to be represented in chemistry (92 percent) and behavioral and social sciences (90 percent) and least likely to be represented in the clinical sciences (19 percent). Individuals with M.D.’s are more likely to be represented in the clinical sciences than in other disciplines. The percentage of M.D.’s who pursue clinical science fields as postdoctoral recipients is approximately 67 percent, compared with less than 10 percent for all other disciplines. Individuals with dual degrees are also more highly represented in the clinical sciences.
ORC Macro 72
In what fields were recipients doing research while receiving Kirschstein-NRSA support?
Research doctorates were more highly represented in chemistry and behavioral sciences, while those with professional doctorates and those with dual degrees were more highly represented in the clinical sciences.
Kirschstein-NRSA postdoctoral fellows were more highly represented in chemistry and the behavioral sciences, while trainees were more highly represented in the clinical sciences and other unspecified fields.
Those who had both Kirschstein-NRSA traineeships and fellowships were more likely to be found in the biomedical sciences.
Figure 3.6. Percentage of Kirschstein-NRSA Postdoctoral Recipients,by Research Area and Type of Degree
Sources: CGAF, TFFNote: Further details are presented in appendix table B.3.
Representation within the different broad fields of study during postdoctoral tenure by type of Kirschstein-NRSA support is presented in figure 3.7. We found that Kirschstein-NRSA fellows are proportionally more represented in chemistry, behavioral and social sciences, and biomedical sciences than trainees, who are proportionally more involved in clinical sciences and other unspecified fields of study. The use of the different postdoctoral support mechanisms for supporting research in different fields of science is apparent from these data. However, to some extent they also reflect the relative use of these mechanisms by those with research and professional doctorates. In other words, there seems to be a relationship that matches certain types of degrees with certain types of support across the fields of study.
ORC Macro 73
Figure 3.7. Percentage of Kirschstein-NRSA Postdoctoral Recipients,by Research Area and Type of Kirschstein-NRSA Postdoctoral Support
Sources: CGAF, TFF Note: Further details are presented in appendix table B.3.
3.4. WHAT WAS THE RESEARCH INTENSITY OF THE INSTITUTION AT WHICH THE KIRSCHSTEIN-NRSA FELLOW OR TRAINEE CONDUCTED RESEARCH?
The choice of institution at which an individual pursues a postdoctoral research position reflects a number of factors, including the mentoring and research opportunities available. Such factors would be expected to assist individuals in obtaining quality research positions after completing their postdoctoral appointments. In figures 3.8 and 3.9, we compare postdoctoral institutions on the basis of their research intensity for Kirschstein-NRSA fellows, trainees, and those who received both fellowships and traineeships. Figure 3.8 compares institutions by whether the institution is designated as a research institution by the Carnegie Classification, and figure 3.9 compares institutions by NIH research funding. Both measures reflect institutional rankings at the time that the individuals performed their postdoctoral work.
ORC Macro 74
What was the research intensity of the institution at which the Kirschstein-NRSA fellow or trainee conducted research?
About three out of every five Kirschstein-NRSA postdoctoral recipients perform their postdoctoral work in a research-intensive institution as defined by the Carnegie Classification.
About four out of every five Kirschstein-NRSA postdoctoral recipients perform their postdoctoral work in an institution ranked in the top 100 in terms of NIH research funding.
Kirschstein-NRSA fellows tend to be relatively less prevalent than trainees in institutions that rank in the top 100 (and top 25) for NIH research funding.
Approximately 63 percent of individuals receiving Kirschstein-NRSA postdoctoral support are at Research I and Research II universities (see figure 3.8). Other individuals perform their postdoctoral studies at universities that have smaller doctoral programs or do not have a doctoral program. It should be noted that less than 3 percent of the postdoctorates are at other doctorate-granting universities, and more than 35 percent are at other unclassified institutions. In general, there are small differences when comparing individuals by type of support on this measure.
Figure 3.8. Percentage of Kirschstein-NRSA Fellows and Trainees Performing Research at Research I or Research II Universities (Carnegie Classification),
by Type of Postdoctoral Support
Sources: CGAF, TFFNote: Further details are presented in appendix table B.4.
Another indicator related to describing the research orientation of the institution involves the amount of grant-related research support that the institution receives from NIH. Figure 3.9 shows the percentage of Kirschstein-NRSA postdoctorates performing their postdoctoral work at the top 100 institutions in terms of NIH research funding. About 84 percent of all Kirschstein-NRSA postdoctoral support recipients perform their postdoctoral work at the top 100 institutions receiving NIH research funding.
Unlike the analysis using the Carnegie Classification, figure 3.9 shows notable differences between individuals with fellowships, traineeships, and those with both fellowships and traineeships. When compared with fellows, trainees are more likely to perform their postdoctoral studies at the institutions that receive the most NIH research funding. Eighty-seven percent of trainees (and individuals with both traineeship and fellowship support) perform their postdoctoral work in the top 100 institutions, compared with 75 percent of Kirschstein-NRSA fellows. This difference is even more dramatic if we consider postdoctoral work performed at the top 25 institutions, with 54 percent of trainees and 38 percent of fellows performing their
ORC Macro 75
postdoctoral work at these institutions. Thus, it would appear that fellows more than trainees tend to work in institutions that are not as well funded by NIH.
Figure 3.9. Percentage of Kirschstein-NRSA Postdoctoral Support Recipients Performing Their Postdoctoral Work at the Top 100 Institutions Receiving NIH Extramural Support,*
by Type of Postdoctoral Support
* At the time of the postdoctoral supportSources: CGAF, TFFNote: Further details are presented in appendix table B.5.
3.5. SUMMARY AND CONCLUSIONS
This chapter focused on characterizing Kirschstein-NRSA postdoctoral training, specifically with regard to the following:
The use of particular Kirschstein-NRSA mechanisms to continue postdoctoral training The length of Kirschstein-NRSA postdoctoral support The research fields in which individuals were pursuing postdoctoral research The research intensity of the institutions in which they were continuing their postdoctoral
research
The findings indicate that, with regard to application, award, and appointee patterns:
Trainees outnumbered fellows by a 3-to-1 margin when examining the first mechanism of support received. Fellows were more likely to have research doctorates than professional degrees. The ratio of trainees to fellows who had research doctorates was 1.9 to 1; the trainee to fellow ratio for M.D.’s was about 2.2 to 1.
About 37 percent of all first-time Kirschstein-NRSA postdoctoral fellowship applicants (excluding those who previously had a traineeship) received awards. M.D./Ph.D.’s had an award rate of 57 percent.
ORC Macro 76
Almost 14 percent of trainees without any previous Kirschstein-NRSA support applied for fellowships, with a 41-percent award rate. Trainees with research doctorates applied at a higher rate than those with professional doctorates and dual degrees, but were less likely to obtain funding through the fellowship mechanism. About 48 percent of these trainees received additional traineeship support.
Twenty-seven percent of individuals who never received Kirschstein-NRSA postdoctoral funding and failed to obtain fellowship funding during their first attempt reapplied for fellowship funding. Of this group, about 27 percent were successful.
With regard to length of support, about 50 percent of Kirschstein-NRSA fellows and trainees received 2 years of support, while about 50 percent of those with both fellowships and traineeships received about 3 years of support.
With regard to field of study, research doctorates were more highly represented in chemistry and behavioral sciences, while those with professional doctorates were highly represented in the clinical sciences. Postdoctoral fellows were more highly represented in chemistry and behavioral sciences. The evidence seems to indicate a segmentation that separates the interests of research and professional doctorates and leads them to different support mechanisms.
Finally, it seems that the research intensity of the institution in which the postdoctoral work is performed (as measured by NIH funding) might be a critical differentiator within the postdoctoral population. About three out of every five Kirschstein-NRSA postdoctoral recipients perform their work in a research-intensive institution as defined by the Carnegie Classification. Differences among Kirschstein-NRSA postdoctoral fellows and trainees are minimal. However, when institutions are categorized by whether they rank in the top 100 in NIH research funding, we find that trainees are more likely to be in the highest-ranked institutions than fellows.
This chapter sought to determine what differences, if any, distinguish various Kirschstein-NRSA postdoctoral populations. The conclusion is that the postdoctoral experience is varied to the extent that it would be difficult to attribute any outcome to a “generic” Kirschstein-NRSA postdoctoral award. Rather, the population is segmented by length of support and by the various mechanisms that are used to sustain individuals through their postdoctoral experience.
ORC Macro 77
4. CAREER OUTCOMES OF KIRSCHSTEIN-NRSA POSTDOCTORAL RECIPIENTS
4.1. INTRODUCTION
A major premise of the Kirschstein-NRSA program is that it is an investment in developing the pool of scientists who can advance the Nation’s research agenda in the biomedical and behavioral science fields. In part, the success of this program can be examined in terms of various research-related outcomes, both in terms of increasing the overall development of scientific capacity and the productivity and success of scientists who have participated in the program.
Previous studies that have assessed the success of Kirschstein-NRSA postdoctoral support have largely focused on the following three types of outcomes:
Success in obtaining NIH research funding. NIH offers a comprehensive grant program to provide funding for research in the biomedical and behavioral sciences. Grant applications are reviewed through a peer-review process, and successful applications signify an area where a potential research contribution is expected to occur. Since the NIH research budget supplies a large portion of biomedical research funding, this outcome would provide a signal of successful outcomes relevant to the Kirschstein-NRSA program. However, this should not suggest that NIH funding is the only successful funding outcome; there are many Federal agencies and private organizations that fund research in these fields.
Publications. The number of publications and number of times those publications are cited are often regarded as an important measure of scientific productivity. Publications, especially those in peer-reviewed journals, represent participation in scholarly discourse and the advancement of scientific knowledge. Citations reflect the degree to which publications have an impact on the scientific community.
Research-focused employment. Employment in research-intensive settings is generally considered a positive outcome of Kirschstein-NRSA postdoctoral support. Tenured employment in academic settings, where much of the independent research occurs, and academic rank are also indicators of the future success of the postdoctoral recipient.
The findings of Garrison and Brown on the effects of postdoctoral training have supported the conclusion that individuals receiving support from Kirschstein-NRSA demonstrate a greater likelihood of applying for and receiving research funding, a higher publication rate, and a greater likelihood of being employed in senior-level positions in academia and other research-focused organizations. In their study, Garrison and Brown compared individuals with Kirschstein-NRSA postdoctoral funding with other doctorate recipients—a population that included individuals dedicated to research careers as well as those with nonresearch careers. With regard to research doctorates, the comparison groups were those individuals who expressed interest in pursuing a postdoctoral position and those who did not.
ORC Macro 78
In this study we have focused on a number of comparisons to more fully explore the effect of the Kirschstein-NRSA postdoctoral program and experience. Our analysis will use the following study groups:
Doctorates who received a Kirschstein-NRSA fellowship through an F32 or F35 grant mechanism. This group includes individuals who may have applied several times prior to being awarded a fellowship and excludes Kirschstein-NRSA fellows who were appointed to Kirschstein-NRSA traineeships either before or after their fellowships.
Doctorates who were appointed to a Kirschstein-NRSA traineeship by an institution with a T32 (training) grant. This group excludes individuals who received both a traineeship and a fellowship during their postdoctoral tenure.
Doctorates who received both Kirschstein-NRSA fellowship and traineeship support. Since this group may contain characteristics of both fellows and trainees, segregating it in our analysis may eliminate any ambiguities concerning comparisons between fellows and trainees.
Doctorates who applied for but did not receive a Kirschstein-NRSA fellowship. This group includes individuals who may have received postdoctoral funding from other sources.
Other non-Kirschstein-NRSA postdoctorates. This group includes research doctorates who were appointed to postdoctoral positions supported by non-Kirschstein-NRSA sources such as NSF. It excludes individuals who applied for Kirschstein-NRSA fellowships and those who were appointed to Kirschstein-NRSA traineeships.
Other research doctorates. This group includes research doctorates who had no indicated postdoctoral activity and did not apply for a Kirschstein-NRSA fellowship.
It should be noted that the first four groups contain individuals who applied for a Kirschstein-NRSA fellowship or received a traineeship appointment, regardless of whether they received research or professional doctorates. The groups constitute the entire population of those who applied for Kirschstein-NRSA fellowships or received Kirschstein-NRSA traineeship appointments. The latter two groups contain only research doctorates because data sources on those with professional degrees who did not receive Kirschstein-NRSA postdoctoral support were not available for this study. The data for these two groups are derived from a sample of individuals who responded to the 1995 SDR.
By using these comparison groups, we hope to address the many nuances of the Kirschstein-NRSA program. For example, we believe that differences in outcomes between Kirschstein-NRSA fellows and trainees may reflect the kind of training that each group receives. The two groups of individuals who did not receive Kirschstein-NRSA postdoctoral support will also provide important comparisons, since these groups contain individuals who have pursued postdoctoral careers with non-Kirschstein-NRSA funding and assumed employment in administrative and nonresearch settings. Thus, comparing outcomes for Kirschstein-NRSA fellows and trainees with those for individuals who have non-Kirschstein-NRSA postdoctoral
ORC Macro 79
positions or have applied unsuccessfully for Kirschstein-NRSA fellowships may provide for more meaningful analysis.
This report largely focuses on presenting outcomes data in the form of percentage comparisons among the study groups. We also compare the study groups using regression analysis. These comparisons are intended to provide evidence of differential outcomes among the study groups, controlling for factors other than study group affiliation. Each of the regression tables presented in this chapter focuses only on the estimates generated for the study groups. The estimates presented should be interpreted as scaled values representing the study groups’ effect when compared with a benchmark study group. For the analysis of research doctorate outcomes, the benchmark study group is other research doctorates. Thus the estimate for other research doctorates will be zero; estimates for all other study groups will represent deviations from zero. The estimates provide a scaled value and generally cannot be interpreted directly. Significance of these deviations from the zero value is shown. Therefore, an indication of significance means that a study group has an outcome that is notably different from the other research doctorates study group. Because the other research doctorates study group may not provide a useful contrast, we also present results that contrast outcomes for one study group (Kirschstein-NRSA trainees) with another study group (Kirschstein-NRSA fellows).
The following sections will present a descriptive analysis of outcomes for NIH research funding, publications, and employment, followed by a statistical analysis that controls for various factors that might affect outcomes independent of Kirschstein-NRSA postdoctoral support.
4.2. NIH RESEARCH FUNDING
Ultimately, NIH traineeships and fellowships are intended to cultivate the ability of supported individuals to pursue independent research. NIH is a substantial source of funding for such research, particularly in the biomedical and behavioral sciences. Our analysis of this outcome measure will focus on a number of outcome measures related to NIH grant award mechanisms, including the following:
Research Project Grant Program (R01). These are awards to universities or institutions that allow investigators to pursue research that was self-initiated or initiated in response to a program announcement. We will examine these grants together with the New Investigator Research Award (NIRA) and First Independent Research Support and Transition (FIRST) Award, which were specifically established to facilitate awards to first-time investigators before they compete for R01 awards. NIRA (R23) was implemented in 1977 to provide funding for individuals with little previous experience as principal investigators. The FIRST (R29) award mechanism superseded NIRA in 1986. As a result of an analysis by the Working Group for New Investigators, these awards were discontinued in favor of allowing individuals to apply through the R01 grant mechanism. However, when NIRA and FIRST existed, they provided alternatives to R01 grants, and for the purposes of this study, we will consider them as equivalent to the R01 grant mechanism.
ORC Macro 80
Other research grant mechanisms are:
Small Grant Program (R03)—Time-limited and nonrenewable grants in targeted areas
Academic Research Enhancement Award (AREA) Grants (R15)—Small-scale projects at institutions that do not receive substantial NIH funding
Exploratory/Developmental Research Grant (R21)—Small, time-limited grant to explore new ideas in targeted areas
Exploratory/Developmental Research Grant, Phase II (R33)—The next level for an R21 project
Method to Extend Research in Time (MERIT) Award (R37)—Provides long-term grant support to investigators whose research competence and productivity are distinctly superior and who are highly likely to continue to perform in an outstanding manner
Career development awards (K awards). These awards are intended to foster the development of individual scientists by providing them with support to focus intensively on research or develop skills in new areas of research or with assistance in furthering their careers. Over the years there have been K award mechanisms with varying objectives. These mechanisms can be described by whether they are aimed at research or professional doctorates and by whether they provide support to young scientists who have not established a track record (early career) or established scientists (mid-career or late career).
The total portfolio of awards available (including those listed above) is referred to as KRUMPT. These include:
Research career program awards/grants Research projects Cooperative agreements Program projects and centers Research training
We will focus mainly on the first category of awards, R01. R01 awards and their equivalents account for the largest portion of the NIH extramural grant budget and are viewed as the primary vehicle for supporting independent research.
4.2.1. Grant-Related Outcome Measures
In the descriptive analysis, we focus on the following outcome measures:
Application rate—Whether an individual seeks NIH funding is an important indicator of an attempt to enter the field of biomedical research. We define the application rate as the number of biomedical and behavioral research doctorates who submitted a competing application to NIH after receiving their doctorates.
ORC Macro 81
Award outcomes—There are three measures within this category:
Percentage of those applying who receive at least one award. This measure attempts to show the prevalence of awards among those who receive doctorates.
Number of awards after receipt of the doctorate Percentage of total applications that resulted in awards, where a particular individual can
have any number of applications. We derive this percentage for each individual in the population and then average the percentage across the whole population.
Mean years until first award—This measure examines the mean number of years that it takes to obtain the first competing award. It indicates how quickly different groups reach principal investigator status.
4.2.2. Data Sources and Methodology
The analysis in this section is in two parts. The first focuses on research doctorates while the second focuses on all doctorates. This approach was adopted for two reasons. First, background data on those with only professional degrees were not available for the study, thus limiting the analysis of those with professional degrees to the information available from the CGAF. Second, and more importantly, comparisons between the study groups needed a reference point that would provide an equal measure of time for individuals to apply for and successfully obtain grant funding. We considered three possible reference points:
Year in which the doctorate was awarded Year in which the individual first received Kirschstein-NRSA postdoctoral support Year in which the individual last received Kirschstein-NRSA postdoctoral support
The year in which the doctorate was awarded is the most appropriate reference point for comparing Kirschstein-NRSA postdoctorates with other research doctorates, at least for those who received their research doctorate prior to continuing with a postdoctorate or pursuing employment or other career paths. Using this reference point is similar to examining outcomes had the Kirschstein-NRSA postdoctoral option not been available.60 For individuals with professional doctorates, it is less clear that this is a satisfactory reference point because those earning professional doctorates may delay a postdoctorate until internships or residencies are completed. In other words, even though professional doctorates may have an interest in pursuing a research career, they may still be pursuing postdoctoral studies long after their professional doctorates were awarded.
It is also clear that the latter two reference points may be the most useful for examining differences among those who receive Kirschstein-NRSA support. Of these two points, the year in which the individual last received Kirschstein-NRSA postdoctoral support is more appropriate because it indicates the year that the individual began his or her career after Kirschstein-NRSA
60 It can also be argued that individuals who do not receive postdoctoral support is a poor counterfactual because of the selection process inherent in choosing Kirschstein-NRSA fellows and trainees; in other words, it does not indicate outcomes achieved if Kirschstein-NRSA did not exist because this is a different population.
ORC Macro 82
support. This reference point does, however, tend to ignore activity during their postdoctoral tenure.
Using the first and third reference points, we constructed two analyses. The first analysis will concentrate on comparisons between Kirschstein-NRSA postdoctorates and other research doctorates using the first reference point. The second analysis consists of the entire Kirschstein-NRSA population, including research and professional doctorates. This analysis uses the third reference point and does not consider those without Kirschstein-NRSA funding. Within each analysis, we will use both descriptive statistics and regression modeling to provide an overall perspective on differences among the groups. The regression modeling is intended to provide information on whether the difference among groups is significant, controlling for a variety of factors, including gender and measures related to the doctoral institution and experience of the individual, among others. It should be noted that one of the factors included in the regression approach is the reference point (year of receipt of doctorate or last year of postdoctoral support), which provides control over the effect of length of professional career on outcomes. In the descriptive analyses, we limit outcomes to the first 10 years after the reference point for ease of presentation.
4.2.3. The Effect of the Kirschstein-NRSA Program for Research Postdoctorates
In comparing research doctorates with Kirschstein-NRSA support to other research doctorates, there are two caveats to keep in mind. First, the Kirschstein-NRSA fellowship and training programs represent a very diverse set of experiences; there may be variations in the outcomes within that population depending on the length of support, type of support received, and kinds of activities in which these individuals were involved. Second, there may be even larger variations among research doctorates who received no support. This population could include those who had no plans to pursue research careers as well as those who obtained postdoctoral funding through non-Kirschstein-NRSA mechanisms. We will first compare all Kirschstein-NRSA postdoctorates to all other research doctorates and then identify grant-related outcomes among the various groups specified in the introduction to this chapter.
ORC Macro 83
4.2.3.1. Application and Award Rates
Figure 4.1 shows the percentage of research doctorates who applied for NIH funding through the following grant mechanisms:
R01 grants Combined R01, R23, and R29 grants
(R01 equivalents) All research project grants KRUMPT grants
The data indicate a relatively large difference in application rates for the R01 mechanism between those who received Kirschstein-NRSA postdoctoral support and those who did not (14 percent in favor of those who received Kirschstein-NRSA support), and even larger margins for the other grant mechanisms. Approximately 36 percent of the Kirschstein-NRSA postdoctoral recipients applied for R01 grants compared with 22 percent of the other research doctorates. This difference increases among the other more inclusive types of grant categories displayed. The higher application rate for these more inclusive categories may indicate that Kirschstein-NRSA postdoctorates were more familiar with award mechanisms other than R01s and thus were more likely to apply for them.
Figure 4.2, however, shows that 54 percent of the Kirschstein-NRSA postdoctorates applying for an R01 grant receive at least one grant. Of other non-Kirschstein-NRSA postdoctorates, 51 percent receive at least one grant. This is a marginal difference at most, but the difference increases slightly as the grant mechanism categories become more inclusive. When all mechanisms are considered (e.g., KRUMPT awards), the difference is approximately 6 percent. It may be the case that Kirschstein-NRSA postdoctoral recipients are not only more familiar with the other non-R01 award mechanisms but also submit applications that that adhere more closely to the objectives of these other mechanisms than non-Kirschstein-NRSA postdoctorates and thus have a higher success rate in obtaining support.
ORC Macro 84
With regard to the percentage of the study population applying for R01 grants and the equivalents, the data indicate the following:
A higher percentage of Kirschstein-NRSA postdoctoral recipients applied for R01 grants than did other study groups.
A higher percentage of Kirschstein-NRSA fellows applied for and received R01 grants than did other study groups.
Although a higher percentage of those with Kirschstein-NRSA traineeships applied for R01 grants, a smaller percentage obtained R01 funding as compared with unsuccessful fellowship applicants and non-Kirschstein-NRSA postdoctorates.
Figure 4.1. Percentage of Research Doctorates Applying for Specified Competing NIH Grants in the First 10 Years After Receipt of the Doctorate, by Kirschstein-NRSA
Postdoctoral Status
Sources: CGAF, TFF
Figure 4.2. Percentage of Research Doctorates Applying for Specified Competing NIH Grants Who Received NIH Grant Funding in the First 10 Years After Receipt of the
Doctorate, by Kirschstein-NRSA Postdoctoral Status
Sources: CGAF, TFF
The next figures present application (figure 4.3) and award (figure 4.4) data for each of the six study groups for the R01 and combined R01 equivalents categories. We focus on these categories since they make up the majority of NIH research funding. Data on R01 grant applications indicate that the highest percentage of applicants applying for an R01 is found
ORC Macro 85
among Kirschstein-NRSA fellows and trainee/fellows (45 and 46 percent, respectively), whereas the lowest percentage is found among other research doctorates (9 percent). Fewer Kirschstein-NRSA trainees than Kirschstein-NRSA fellows apply for R01 support (30 percent). All the percentages increase when the R23 and R29 mechanisms are factored in, but the percentage increase is notably larger among Kirschstein-NRSA postdoctorates than those without such support.
Figure 4.4 indicates that of those individuals applying for an R01, Kirschstein-NRSA fellows are more likely to be awarded at least one award than any other group. Other doctorates are almost half as likely to succeed in obtaining R01 funding. It is important to note that only 47 percent of Kirschstein-NRSA trainees received funding within 10 years of their doctorate, compared with 54 percent of unsuccessful fellowship applicants and 49 percent of other non-Kirschstein-NRSA postdoctorates.
Figure 4.3. Percentage of Research Doctorates Applying for Competing R01 and R01-Equivalent NIH Grant Funding in the First 10 Years After Receipt of the Doctorate,
by Study Group
Sources: CGAF, TFFNote: Further details are presented in in appendix table C.1.
ORC Macro 86
Figure 4.4. Percentage of Research Doctorates Receiving Competing R01 and R01-Equivalent NIH Grant Funding of Those Applying in the First 10 Years After Receipt
of the Doctorate, by Study Group
Sources: CGAF, TFFNote: Further details are presented in in appendix table C.2.
Table 4.1 presents the results of the logistic regression that models the probability of applying for a competing R01 grant only or for competing R01-equivalent grants. The parameter estimates indicate deviations from the mean probability exhibited in the population (using other research doctorates as the baseline group). It should also be noted that unlike the data presented in figures 4.1–4.4, which examine the 10-year period after the doctorate, these data consider the application’s history from receipt of the doctorate to 2004. Since some individuals could have up to 40 years of experience after the doctorate and others may have as few as 12 years, we included a term to control for this varying length of experience. The regressions, as do all other regressions presented in this chapter, control for a variety of other factors that may describe differences in the population.
The results from the logistic regression analysis indicate that the probability of applying for the R01 and combined R01-equivalent category is highest among Kirschstein-NRSA postdoctoral support recipients. This generally supports the data presented in figure 4.2, although there is no significant difference between Kirschstein-NRSA trainees and unsuccessful fellowship applicants exhibited here for R01s. It should be noted that while the parameter estimates increase for Kirschstein-NRSA postdoctorates for R01 equivalents, they remain approximately the same
ORC Macro 87
for the other groups. This reinforces the notion that the R23 and R29 mechanisms were more familiar to those involved in the Kirschstein-NRSA program.
Table 4.1. Regression Results Reflecting the Likelihood of Individuals Submitting a Competing R01 or Combined R01-Equivalent Application
GroupParameter Estimate
R01s1 R01 Equivalents2
Kirschstein-NRSA trainee/fellows 2.02* 2.15*Kirschstein-NRSA fellows 1.93* 2.01*Kirschstein-NRSA trainees 1.38* 1.50*Unsuccessful fellowship applicants 1.38* 1.42*Other non-Kirschstein-NRSA postdoctorates 0.67* 0.67*Other research doctorates 0.00 0.00
ContrastsFellows versus trainees Significant (application rate
higher for fellows)Significant (application rate higher for fellows)
Unsuccessful fellowship applicants versus trainees
Not significant Significant (application rate higher for trainees)
Trainees versus non-Kirschstein-NRSA postdoctorates
Significant (application rate higher for trainees)
Significant (application rate higher for trainees)
Number of observations 39,045 39,0451 Regression data can be found in appendix table D.1.2 Regression data can be found in appendix table D.2.* Indicates statistical significance at p = 0.05
Table 4.2 presents the regression results of being awarded at least one competing grant. The results here are based only on those submitting a competing application. The data show that Kirschstein-NRSA fellows and trainee/fellows have greater success than all the other study groups, including unsuccessful fellowship applicants. The data also indicate that Kirschstein-NRSA trainees are significantly less likely to obtain at least one award than unsuccessful fellowship applicants. The difference between trainees and other non-NRSA postdoctorates is not significant. The ordering of the parameter estimates is not affected much when R23s and R29s are factored into the calculation.
ORC Macro 88
Table 4.2. Regression Results Reflecting the Likelihood of a Competing R01 or R01-Equivalent Award Given the Submission of an Application
GroupParameter Estimate
R01s1 R01 Equivalents2
Kirschstein-NRSA fellows 0.64* 0.70*Kirschstein-NRSA trainee/fellows 0.52* 0.60*Unsuccessful fellowship applicants 0.38* 0.38*Other non-Kirschstein-NRSA postdoctorates 0.21 0.16Kirschstein-NRSA trainees 0.14 0.19*Other research doctorates 0.00 0.00
ContrastsFellows versus trainees Significant (probability of
being awarded a competing R01 is higher for fellows)
Significant (probability of being awarded a competing R01 is higher for fellows)
Unsuccessful fellowship applicants versus trainees Significant (probability of being awarded a competing R01 is higher for unsuccessful fellowship applicants)
Significant (probability of being awarded a competing R01 is higher for unsuccessful fellowship applicants)
Trainees versus other non-Kirschstein-NRSA postdoctorates
Not significant Not significant
Number of observations 14,357 15,8251 Regression data can be found in appendix table D.3.2 Regression data can be found in appendix table D.4. * Indicates statistical significance at p = 0.05
4.2.3.2. Number of Awards
The number of grant awards is an indicator of the repeated success of individuals in obtaining research funding through R01 grants and other mechanisms. The distribution of awards shows an overall difference among the study groups. Table 4.3 presents information on the distribution of R01 awards by group within the first 10 years after receipt of the doctorate. The table only presents data on those receiving at least one R01 award.
For all groups, there is a positive skew in the distributions. This positive skew is most evident for Kirschstein-NRSA trainees and for other research doctorates, meaning that they are less likely to obtain multiple R01 grants. On the other hand, individuals with four or more awards are more prevalent among Kirschstein-NRSA fellows (6 percent), unsuccessful Kirschstein-NRSA fellowship applicants (6 percent), and other non-Kirschstein-NRSA postdoctoral support recipients (7 percent). The higher percentages in the latter two groups may be attributable to individuals who assumed research positions that allowed them to pursue grant support immediately after their doctorate.
ORC Macro 89
Analysis of the distribution of R01 awards indicates the following:
Kirschstein-NRSA fellows receive a greater number of awards.
Unsuccessful fellowship applicants and Kirschstein-NRSA trainee/fellows receive the second highest number of awards.
Kirschstein-NRSA trainees receive as many awards as other non-Kirschstein-NRSA postdoctorates.
TABLE 4.3. DISTRIBUTION OF STUDY GROUP RESEARCH DOCTORATES, BY NUMBER OF R01 AWARDS
Study GroupPercentage of Research Doctorates With Awards
1 Award 2 Awards 3 Awards 4 or More Awards Total
Kirschstein-NRSA fellows 51.7 30.9 11.8 5.6 100.0
Kirschstein-NRSA trainees 63.4 24.9 8.6 3.0 100.0
Kirschstein-NRSA trainee/fellows 56.3 29.7 10.2 3.9 100.0
Unsuccessful fellowship applicants 53.8 29.0 11.4 5.8 100.0
Other non-Kirschstein-NRSA postdoctorates
51.2 29.2 13.0 6.7 100.0
Other research doctorates 68.1 20.9 7.5 3.6 100.0
The data presented in Table 4.4 were modeled through a Poisson regression with a dependent variable representing the number of awards obtained. Again, unlike the data presented in table 4.3, the regression considers the entire career of the study groups. The regression parameters indicate that the most successful group was Kirschstein-NRSA fellows. Kirschstein-NRSA trainees, on the other hand, did not secure as many competing R01 grants as unsuccessful fellowship applicants and did not distinguish themselves from other non-Kirschstein-NRSA postdoctoral support recipients. As we would expect, the addition of R23 and R29 awards did not change the parameters in any notable way. (These one-time awards would have at best contributed one grant to a recipient’s total number of grants.)
The results in the regression, when looked at in terms of the data presented in table 4.3, may also suggest that over the course of a career Kirschstein-NRSA fellows and trainee/fellows outperform unsuccessful fellowship applicants and other postdoctorates (although this trend is not as evident over the short run). The same trend can be identified for trainees, to a lesser extent.
ORC Macro 90
Table 4.4. Regression Results Relating to the Number of Competing R01 or R01-Equivalent Awards
GroupParameter Estimate
R01s1 R01 Equivalents2
Kirschstein-NRSA fellows 0.45* 0.44*Kirschstein-NRSA trainee/fellows 0.38* 0.39*Unsuccessful fellowship applicants 0.33* 0.31*
Other non-Kirschstein-NRSA postdoctorates 0.17* 0.17*Kirschstein-NRSA trainees 0.13* 0.13*Other research doctorates 0.00 0.00
ContrastsFellows versus trainees Significant (fellows are
awarded more R01s)Significant (fellows are awarded more R01s)
Unsuccessful fellowship applicants versus trainees Significant (unsuccessful fellowship applicants are awarded more R01s)
Significant (unsuccessful fellowship applicants are awarded more R01s)
Trainees versus other non-Kirschstein-NRSA postdoctorates
Not significant Not significant
Number of observations 14,357 15,8251 Regression data can be found in appendix table D.5.2 Regression data can be found in appendix table D.6. * Indicates statistical significance at p = 0.05
4.2.3.3. Award Rate
Another measure of success that focuses on individual outcomes is the average individual award rate for competing grants. This measure takes into account the success rate of individuals over all the grant applications submitted. As figure 4.5 illustrates, 34 percent of R01 applications by Kirschstein-NRSA fellows are funded. Other research doctorates are only awarded a grant in 19 percent of their applications—less than half the rate of fellows. The other groups range from 25 to 30 percent, including Kirschstein-NRSA trainees (25 percent).
We do not see a large difference when R23 and R29 grant awards are included primarily because these mechanisms were not widely used and typically resulted in only one award when they were.
ORC Macro 91
Analysis of the average award rate for individuals indicates the following:
Kirschstein-NRSA fellows are most successful. Kirschstein-NRSA trainees are least successful, with
an equivalent rate to other research doctorates.
Figure 4.5. Average Award Rate Among Research Doctorates Receiving Competing R01 and R01-Equivalent NIH Grant Funding in the First 10 Years After Receipt of the
Doctorate, by Study Group
Sources: CGAF, TFFNote: Further details are presented in in appendix table C.3.
Table 4.5 presents the results of a regression analysis in which the dependent variable is the number of awards over the number of applications submitted (in other words, it is the award rate for the individual). Again, the regression considers the whole span of the researcher’s career as opposed to only the 10 years after award of the doctorate as in figure 4.5. The regression shows that Kirschstein-NRSA fellows have the highest award rate; Kirschstein-NRSA trainees have the lowest award rate.61 It should also be noted that the difference between other postdoctoral support recipients and Kirschstein-NRSA trainees has increased to the point of being statistically significant. This difference disappears when R23s and R29s are considered, meaning that trainees enjoyed a higher R23 or R29 award rate.
61 The rankings in the regression are different than those presented in figure 4.5. This difference may reflect the timespan referenced in each analysis. It might suggest, for instance, that trainees initially have more success relative to the baseline group, but the success declines with time. It may also indicate differences in the characteristics of the population that are controlled for by the regression.
ORC Macro 92
Table 4.5. Regression Results Relating to the Individual Average Award Rate for Competing R01 or R01-Equivalent Awards
GroupParameter Estimate
R01s1 R01 Equivalents2
Kirschstein-NRSA fellows 0.06* 0.07*Unsuccessful fellowship applicants 0.03* 0.04*Kirschstein-NRSA trainee/fellows 0.04* 0.04*Other Kirschstein-NRSA postdoctorates 0.03 0.02Other research doctorates 0.00 0.00Kirschstein-NRSA trainees -0.02 0.02
ContrastsFellows versus trainees Significant (fellows have a
higher percentage of R01s awarded)
Significant (fellows have a higher percentage of R01s awarded)
Unsuccessful fellowship applicants versus trainees Significant (unsuccessful fellowship applicants have a higher percentage of R01s awarded)
Significant (unsuccessful fellowship applicants have a higher percentage of R01s awarded)
Trainees versus other non-Kirschstein-NRSA postdoctorates
Significant (other non-Kirschstein-NRSA postdoctoral recipients have a higher percentage of R01s awarded)
Not significant
Number of observations 14,357 15,8251 Regression data can be found in appendix table D.7.2 Regression data can be found in appendix table D.8.* Indicates statistical significance at p = 0.05
4.2.3.4. Years to First Competing R01 or R01-Equivalent Award
The final measure relates to the time elapsed between the award of the doctorate and the first research grant. It must be emphasized that unlike previous tables, the groups analyzed here include only those individuals who were awarded a grant. In this analysis, other research doctorates were awarded an R01 grant soonest (7 years) after their research doctorate. We might expect this since these were individuals in a position to immediately begin searching for and obtaining research funding. It is important to remember that these are successful R01 grant applicants. At the other end of the continuum are Kirschstein-NRSA trainee/fellows (8 years) and Kirschstein-NRSA trainees (9 years). The longer time for the trainee/fellow group is what we expect since these individuals usually have longer postdoctoral tenures. The data indicate that postdoctoral support recipients experience a career lag relative to individuals who enter a research-oriented job directly after receipt of the doctorate. When R01s are combined with R23s and R29s, the period from postdoctorate to first award is shortened by 2 years for Kirschstein-NRSA postdoctorates and 1 year for the other study groups.
ORC Macro 93
With regard to the amount of time it takes to be awarded an R01 after the award of the doctorate, the data indicate the following:
The group that has not had any obvious postdoctoral experience has, on average, the shortest period until they receive an award.
Kirschstein-NRSA fellows take a relatively shorter time to obtain their first R01 grant funding.
Kirschstein-NRSA trainees take the longest time to obtain their first R01 grant funding.
Figure 4.6. Number of Years to Receipt of First Competing R01 and R01-Equivalent Grants in the First 10 Years After Receipt of the Doctorate, by Study Group
Sources: CGAF, TFFNote: Further details are presented in in appendix table C.4.
Table 4.6 presents the results of a regression analysis of the elapsed time between receipt of the doctorate and receipt of the first R01 or R01-equivalent award. The longest time to award was evident for trainees—significantly longer than for most of the other comparison groups. It is notable that trainee/fellows generally spend more time supported by Kirschstein-NRSA and thus probably do not realize their first success until notably after that observed for Kirschstein-NRSA fellows, who tend to have less support. It can be said that the Kirschstein-NRSA delays the award for Kirschstein-NRSA fellows and other groups by more than 1 year over other research doctorates; it delays the award for Kirschstein-NRSA trainees for more than 2 years.
ORC Macro 94
Table 4.6. Regression Results Relating to the Years from Receipt of the Doctorate to First Competing R01 or R01-Equivalent Awards
GroupParameter Estimates
R01s1 R01 Equivalents2
Kirschstein-NRSA trainees 2.05* 1.43*Kirschstein-NRSA trainee/fellows 1.70* 1.08*Other non-Kirschstein-NRSA postdoctorates 1.36* 1.14*Unsuccessful fellowship applicants 1.34* 1.37*Kirschstein-NRSA fellows 1.22* 0.90*Other research doctorates 0.00 0.00
ContrastsFellows versus trainees Significant (fellows are
awarded a first R01 earlier)
Significant (fellows are awarded a first R01 equivalent earlier)
Unsuccessful fellowship applicants versus trainees Significant (trainees are awarded a first R01 later)
Not significant
Trainees versus other non-Kirschstein-NRSA postdoctorates
Significant (trainees are awarded a first R01 later)
Not significant
Number of observations 8,558 9,7891 Regression data can be found in appendix table D.9.2 Regression data can be found in appendix table D.10.* Indicates statistical significance at p = 0.05
4.2.3.5. Summary
In terms of research grant outcomes, the data presented in this section indicate that:
Kirschstein-NRSA fellows, on most measures related to grant outcomes, performed better than most other groups.
Kirschstein-NRSA trainees, in contrast, performed at or below the level of individuals who applied for but did not receive a Kirschstein-NRSA fellowship and other non-Kirschstein-NRSA postdoctorates.
Kirschstein-NRSA trainees who also had a fellowship performed close to fellows on most measures.
Some of the measures seem to indicate that being a Kirschstein-NRSA postdoctoral recipient boosts an individual’s ability to prepare for and receive grants other than R01s.
One measure (years to first R01 award) seems to indicate a lag for Kirschstein-NRSA postdoctoral recipients when compared with research doctorates who presumably become employed immediately after completion of the doctorate.
4.2.4. The Effects of the Kirschstein-NRSA Fellowship and Traineeship Programs on Research and Professional Doctorates
The last section focused on research doctorates, or Ph.D.’s, and described how the outcomes of Kirschstein-NRSA postdoctorates differed from those of other individuals who had research doctorates. Individuals with M.D.’s were excluded from the analysis, primarily due to a career path that traditionally emphasizes the provision of health care services over research. Because of this emphasis in medical school and because of the subsequent training (internships and residencies) that many M.D.’s undertake, the career path of M.D.’s is somewhat different from
ORC Macro 95
that of Ph.D.’s. Those M.D.’s who are interested in research may not be exactly equivalent to Ph.D.’s in terms of when and how they make career choices and are certainly different from M.D.’s who focus on providing health care services. These factors make it difficult to establish an equivalent comparison group, since the most accessible comparison group members are engaged in providing medical services to patients, not engaging in research.
In this section, we will analyze the NIH research grant outcomes for three groups of Kirschstein-NRSA postdoctorates: Ph.D.’s, M.D.’s, and M.D./Ph.D.’s. Analyses will be conducted within each of the three Kirschstein-NRSA postdoctorate groups that have been previously defined (fellows, trainees, and trainee/fellows). The analysis of Ph.D.’s should produce results similar to those observed in the previous section, except that the observation period starts at the end of the last year of Kirschstein-NRSA support, not at the year the doctorate was awarded. By shifting the reference period forward, we are focusing on outcomes that occur after individuals finish their postdoctoral training, at the point that the Kirschstein-NRSA postdoctorates are presumed ready to participate fully in achieving success on the outcomes we are examining. The analysis, as in the last section, will present the results for the following measures:
Application rate Award outcomes
Percentage of applicants with an award Number of awards Award rate
Mean years to first award after completion of the doctorate
The focus will be on the R01 and R01-equivalent grant mechanisms. However, it may help to place these awards in the context of other research-related awards. One type of award that is particularly relevant for these comparisons is the career development (K) awards, which provide assistance to individuals who need additional mentoring or assistance in establishing their careers or changing the direction of their studies. Some of the K award mechanisms have been geared toward helping M.D.’s establish themselves as researchers. In general, the data show that M.D.’s apply for these awards to a greater extent than the other degree study groups. They are also awarded these grants about twice as often as M.D./Ph.D.’s and four times as often as Ph.D.’s. Specific information on the application and award patterns for these K awards is provided in appendix C.
ORC Macro 96
4.2.4.1. NIH Grant Application and Award Rates
Application and award rates refer to the proportion of the population who apply for a research grant and, of those individuals, the proportion who secure an award. We examine these rates for Kirschstein-NRSA fellows, trainees, and trainee/fellows.
Figure 4.7 shows the percentage of Kirschstein-NRSA postdoctoral fellows applying for various awards within 10 years of their last year of Kirschstein-NRSA postdoctoral support. In the last section, which focused on Ph.D.’s only, we found that the percentage of Kirschstein-NRSA postdoctorates applying for grants increased as the grants became more inclusive. This finding is repeated in this analysis across all degree groups (see figure 4.7). However, the application rate for M.D.’s and M.D./Ph.D.’s is higher than the application rate for Ph.D.’s as the categories become more inclusive. This suggests that these two groups (M.D.’s and M.D./Ph.D.’s) applied for R23s and R29s, in particular. M.D.’s apply for R01s and R01 equivalents at a slightly higher rate than Ph.D.’s (51 and 49 percent, respectively, for R01s and 60 and 54 percent, respectively, for R01 equivalents).
With regard to R01 applications, there was very little difference between the three groups, with about half the Kirschstein-NRSA fellows in each degree group applying. Although a higher percentage of Kirschstein-NRSA postdoctorates with M.D.’s apply for R01s than both M.D./Ph.D.’s and Ph.D.’s, proportionately more Kirschstein-NRSA postdoctorates with M.D./Ph.D.’s and Ph.D.’s are actually awarded R01s than M.D.’s. The difference between M.D.’s and Ph.D.’s for R01s is 8 percent (55 and 63 percent, respectively). This difference increases to 10 percent for R01 equivalents (58 and 68 percent, respectively).
ORC Macro 97
With regard to the percentage of doctorates applying for and receiving R01 and R01-equivalent grants:
Proportionally more fellows than trainees apply for and receive grants.
M.D./Ph.D.’s apply for and are awarded grants to a greater extent than the other degree groups.
M.D.’s apply for grants more frequently than Ph.D.’s, but are equally likely to receive an award.
Figure 4.7. Application Rate of Kirschstein-NRSA Postdoctoral Fellows Within 10 Years of the Last Year of Kirschstein-NRSA Postdoctoral Support
Note: Further details are presented in in appendix table C.5.
Figure 4.8. Award Rate of Kirschstein-NRSA Postdoctoral Fellows Within 10 Years of the Last Year of Kirschstein-NRSA Postdoctoral Support
Note: Further details are presented in in appendix table C.6.
Proportionately more M.D./Ph.D.’s receiving postdoctoral traineeships applied for R01s (43 percent) than M.D.’s and Ph.D.’s (both 33 percent). This pattern is also observed for R01 equivalents and all RPGs, where the difference ranges between 8 and 10 percent (see figure 4.9).
With regard to award rates, M.D./Ph.D.’s again are more successful than either M.D.’s or Ph.D.’s, but to a lesser degree than for application rates (see figure 4.10). For example, M.D./Ph.D.’s are 6 percent more successful at receiving R01 equivalents than Ph.D.’s or M.D.’s.
Figure 4.9. Application Rate of Kirschstein-NRSA Postdoctoral Trainees Within 10 Years of the Last Year of Kirschstein-NRSA Postdoctoral Support
ORC Macro 98
Note: Further details are presented in appendix table C.5.Figure 4.10. Award Rate of Kirschstein-NRSA Postdoctoral Trainees Within 10 Years of
the Last Year of Kirschstein-NRSA Postdoctoral Support
Note: Further details are presented in in appendix table C.6.
For those individuals who received both a Kirschstein-NRSA postdoctoral traineeship and fellowship, the M.D./Ph.D. group applied more often for R01s, R01 equivalents, and all RPGs than both the M.D. and Ph.D. groups (see figure 4.11). With regard to awards, there is very little difference among the three degree groups (see figure 4.12). For example, about 60 percent of the trainee/fellows with M.D./Ph.D.’s, 62 percent of the Ph.D.’s, and 57 percent of the M.D.’s received R01 awards.
Figure 4.11. Application Rate of Kirschstein-NRSA Postdoctoral Trainee/Fellows Within 10 Years of the Last Year of Kirschstein-NRSA Postdoctoral Support
ORC Macro 99
Note: Further details are presented in in appendix table C.5.
Figure 4.12. Award Rates of Kirschstein-NRSA Postdoctoral Trainee/Fellows Within 10 Years of the Last Year of Kirschstein-NRSA Postdoctoral Support
Note: Further details are presented in appendix table C.6.
Tables 4.7 and 4.8 present logistic regression results modeling the main effects of the type of Kirschstein-NRSA postdoctoral support and degree relative to R01 and R01-equivalent applications and awards. Table 4.7 indicates that Kirschstein-NRSA fellows and trainee/fellows apply at a higher rate than Kirschstein-NRSA trainees. In addition, it indicates that M.D./Ph.D.’s and M.D.’s both have a higher application rate than Ph.D.’s, reinforcing the results presented in figures 4.7–4.12.
ORC Macro 100
Table 4.7. Regression Results Reflecting the Likelihood of Kirschstein-NRSA Postdoctorates Submitting a Competing R01 or Combined R01-Equivalent Application
Kirschstein-NRSA Group Parameter EstimateR01s1 R01 Equivalents2
Kirschstein-NRSA fellows 0.00 -0.02Kirschstein-NRSA trainees/fellows 0.00 0.00Kirschstein-NRSA trainees -0.57* -0.58*
Degree GroupM.D. 0.15* 0.20*M.D./Ph.D. 0.37* 0.37*Ph.D. 0.00 0.00
Months of Support0–9 -0.39* -0.5010–12 -0.41* -0.5013–24 -0.17* -0.2125–36 0.05 0.0337+ 0.00 0.00
ContrastsFellows versus trainees Significant (fellows are more likely
to apply for R01s)Significant (fellows are more likely to apply for R01 equivalents)
M.D.’s versus Ph.D.’s Significant (M.D.’s are more likely to apply for R01s)
Significant (M.D.’s are more likely to apply for R01 equivalents)
Number of observations 30,729 30,7291 Regression data can be found in appendix table D.11.2 Regression data can be found in appendix table D.12.* Indicates statistical significance at p = .05
Table 4.8 provides results relating to the likelihood of obtaining at least one award. Results indicate that Kirschstein-NRSA fellows had the highest likelihood of receiving at least one R01 award, and trainees had the lowest likelihood. With regard to degree groups, M.D./Ph.D.’s had the highest probability of receiving an award. There was no significant difference between M.D.’s and Ph.D.’s.
In contrast to the last analysis, which indicated that those with more months of Kirschstein-NRSA support were more likely to apply for an R01, this analysis shows that those with fewer months of support are more likely to receive an R01. Table 4.8 suggests that the likelihood is highest among those with less than 9 months of support, while the likelihood is about the same for those with 13–36 months of support.
ORC Macro 101
Table 4.8. Regression Results Reflecting the Likelihood of Kirschstein-NRSA Postdoctorates Being Awarded a Competing R01 or Combined R01-Equivalent Grant
Kirschstein-NRSA GroupParameter Estimate
R01s1 R01 Equivalents2
Kirschstein-NRSA fellows 0.13* 0.11Kirschstein-NRSA trainee/fellows 0.00 0.00Kirschstein-NRSA trainees -0.38* -0.41*
Degree Group
M.D./Ph.D. 0.11 0.13*Ph.D. 0.00 0.00M.D. -0.05 -0.09*
Months of Support0–9 0.11 0.0510–12 0.00 -0.0513–24 0.04 0.0325–36 0.05 0.0537+ 0.00 0.00
ContrastsFellows versus trainees Significant (fellows are
more likely to receive R01s)
Significant (fellows are more likely to receive R01 equivalents)
M.D.’s versus Ph.D.’s Not significant Not significantNumber of observations 13,581 15,0441 Regression data can be found in appendix table D.13.2 Regression data can be found in appendix table D.14.* Indicates statistical significance at p = .05
4.2.4.2. Number of Awards
The discussion thus far considers whether Kirschstein-NRSA postdoctorates received at least one R01 or R01-equivalent award, not the total number of awards received. An analysis of the distribution of the number of R01 awards provides more information than an analysis of the mean or median. Of those receiving awards (table 4.9), 52 percent of all Kirschstein-NRSA postdoctorates, regardless of type of degree and type of Kirschstein-NRSA postdoctoral support, receive at least two awards, and 9 percent receive four or more awards within 10 years of finishing their postdoctorates. Among Kirschstein-NRSA fellows, M.D.’s are less successful than M.D./Ph.D.’s and Ph.D.’s. In particular, M.D./Ph.D.’s secure more R01 awards than the other groups. For Kirschstein-NRSA trainees, M.D./Ph.D.’s are more successful in obtaining R01 grant funding. For example, 26 percent of M.D./Ph.D.’s were awarded three or more awards, compared with 19 percent of Ph.D.’s and 17 percent of M.D.’s. A similar pattern is found among Kirschstein-NRSA trainee/fellows.
ORC Macro 102
With regard to the number of R01 awards:
M.D./Ph.D.’s secure a greater number of grants than the other groups, particularly within the Kirschstein-NRSA trainee group.
Ph.D.’s secure a greater number of grants than M.D.’s.
Table 4.9. Distribution of the Number of R01 Awards, by Study Group and Degree
Kirschstein-NRSA Group Degree
Number of R01 Awards
1 Award 2 Awards 3 Awards 4+ Awards Total
Kirschstein-NRSA fellows
M.D. 50.5% 26.6% 13.3% 9.6% 100.0%M.D./Ph.D. 44.6 24.8 16.6 14.0 100.0Ph.D. 42.4 29.3 16.6 11.7 100.0Total 43.1 29.0 16.3 11.6 100.0
Kirschstein-NRSA trainees
M.D. 54.2 28.7 9.5 7.5 100.0M.D./Ph.D. 48.2 26.0 15.7 10.1 100.0Ph.D. 53.0 27.7 13.1 6.3 100.0Total 52.5 27.9 12.5 7.0 100.0
Kirschstein-NRSA trainee/fellows
M.D. 49.3 28.9 11.2 10.5 100.0M.D./Ph.D. 41.1 26.7 22.2 10.0 100.0Ph.D. 45.8 28.6 15.6 10.0 100.0Total 46.4 28.4 15.3 9.9 100.0
Total 48.2 28.4 14.3 9.1 100.0
This relationship is further explored through Poisson regression of the number of R01 and R01-equivalent grants (see table 4.10). The results indicate that Kirschstein-NRSA trainees are awarded fewer R01 grants, and Kirschstein-NRSA fellows are awarded more than Kirschstein-NRSA trainee/fellows. With regard to degree group, M.D./Ph.D.’s are awarded the most R01 grants, and M.D.’s are awarded the least. The difference between M.D.’s and Ph.D.’s is not significant. When the Kirschstein-NRSA population is compared by months of support, we find that those with 9 or fewer months of support are awarded the greatest number of R01 grants.
ORC Macro 103
Table 4.10. Regression Results of the Number of R01 and R01-Equivalent Awards, Given the Submission of a Competing R01 and R01-Equivalent Application
Kirschstein-NRSA GroupParameter Estimate
R01s1 R01 Equivalent2
Kirschstein-NRSA fellows 0.06* 0.05*Kirschstein-NRSA trainee/fellows 0.00 0.00Kirschstein-NRSA trainees -0.26* -0.27*
Degree GroupM.D./Ph.D. 0.15* 0.14*Ph.D. 0.00 0.0M.D. -0.03* -0.04*
Months of Support0–9 0.11* 0.11*10–12 -0.04 -0.0413–24 -0.02 -0.0225–36 0.02 0.0237+ 0.00 0.00
ContrastsFellows versus trainees Significant (fellows are
awarded more R01s)Significant (fellows are awarded more R01 equivalents)
M.D.’s versus Ph.D.’s Not significant Significant (Ph.D.’s are awarded more R01 equivalents)
Number of observations 13,581 15,0441 Regression data can be found in appendix table D.15.2 Regression data can be found in appendix table D.16.* Indicates statistical significance at p = .05
4.2.4.3. Award Rates
The award rate is defined as the number of awards divided by the number of applications for a person. In figures 4.13, 4.14, and 4.15, we explore award rates within each Kirschstein-NRSA postdoctorate study group by type of degree. Within the Kirschstein-NRSA fellows group (see figure 4.13), the award rate for R01s is highest among Ph.D.’s (33 percent) and second highest among those who have M.D./Ph.D.’s (30 percent). The award rate for M.D.’s is 25 percent.
In contrast, when we consider Kirschstein-NRSA trainees, we find that Ph.D.’s (24 percent) have the same award rate as M.D.’s (25 percent), and both groups have a slightly lower award rate than those with M.D./Ph.D.’s (28 percent) (see figure 4.14).
For those who had both traineeship and fellowship support, all three groups had about the same award rate (see figure 4.15). It should also be noted that when the three Kirschstein-NRSA groups are compared by type of degree, the award rates for M.D./Ph.D.’s are similar across the groups, while Ph.D.’s varied from a high of 31 percent for Kirschstein-NRSA fellows to
ORC Macro 104
With regard to the R01 award rate:
Fellows with Ph.D.’s and M.D./Ph.D.’s have a higher award rate than those with M.D.’s.
There is very little difference in award rate within the trainee and trainee/fellow groups by degree.
24 percent for Kirschstein-NRSA trainees. It would seem that fellows with Ph.D.’s have the highest award rate of all the groups and surpass the performance of trainees with Ph.D.’s.
Figure 4.13. Average Award Rate for R01 and R01-Equivalent Grant Applications Submitted Within 10 Years After Completion of the Kirschstein-NRSA Postdoctorate for
Kirschstein-NRSA Fellows, by Degree
Note: Further details are presented in in appendix table C.7.
Figure 4.14. Average Award Rate for R01 and R01-Equivalent Grant Applications Submitted Within 10 Years After Completion of the Kirschstein-NRSA Postdoctorate for
Kirschstein-NRSA Trainees, by Degree
Note: Further details are presented in appendix table C.7.
ORC Macro 105
Figure 4.15. Average Award Rate for R01 and R01-Equivalent Grant Applications Submitted Within 10 Years After Completion of the Kirschstein-NRSA Postdoctorate for
Kirschstein-NRSA Trainee/Fellows, by Degree
Note: Further details are presented in appendix table C.7.
The regression results for the award rate are presented in table 4.11. This table indicates that Kirschstein-NRSA fellows have the highest success rates, Kirschstein-NRSA trainees have the lowest, and the difference is significant. However, there is no difference among the three degree groups, nor is there any significant difference by months of support. Thus the training mechanism, and not degree, is important in explaining R01 award rate variation. It should be noted that the descriptive analysis above seems to indicate an interaction effect, which was not included in the regression.
ORC Macro 106
Table 4.11. Results From a Regression Analysis of the Average Award Rate (Proportion of Grants Awarded to Grants Competed for Kirschstein-NRSA Postdoctoral Recipients Only) for Competing R01 and R01-Equivalent Grants
Kirschstein-NRSA Group Parameter EstimateR01s1 R01 Equivalents2
Kirschstein-NRSA fellows 0.02* 0.01Kirschstein-NRSA trainee/fellows 0.00 0.00Kirschstein-NRSA trainees -0.04* -0.04*
Degree GroupM.D./Ph.D. 0.01 0.01Ph.D. 0.00 0.00M.D. -0.01 -0.01
Months of Support0–9 0.03 0.0210–12 0.01 0.0013–24 0.00 0.0025–36 0.00 -0.0137+ 0.00 0.00
ContrastsFellows versus trainees Significant (fellows were
more successful in receiving R01s)
Significant (fellows were more successful in receiving R01 equivalents)
M.D.’s versus Ph.D.’s Not significant Not significantNumber of observations 13,581 15,044
1 Regression data can be found in appendix table D.17.2 Regression data can be found in appendix table D.18.* Indicates statistical significance at p = .05
4.2.4.4. Years to First Award
The last measure explored is the length of time it takes for individuals to receive their first R01 or R01-equivalent award after completion of the doctorate. Figures 4.16–4.18 present data for Kirschstein-NRSA fellows, trainees, and trainee/fellows, respectively, by degree. Across all groups, R01-equivalent awards were received 1–1.7 years earlier than R01s. Thus, the R23 and R29 awards effectively reduced time to the first award. The gap varied somewhat across the groups, with the shortest gaps associated with fellows with Ph.D.’s and fellows and trainees with M.D./Ph.D.’s.
With regard to fellows, Ph.D.’s were awarded their first R01 grants about 1.5 years earlier than M.D.’s (see figure 4.16). M.D./Ph.D.’s lagged behind Ph.D.’s by approximately 2 months. The same pattern is evident for trainees, although Kirschstein-NRSA trainees took longer than Kirschstein-NRSA fellows to secure their first R01 funding. Within the trainee group, Ph.D.’s took 6.5 years and M.D.’s took 7.6 years. Finally, for trainee/fellows the pattern is close to that of fellows, except that the gap between Ph.D.’s and M.D.’s is smaller.
Figure 4.16. Years to First Award for R01 and R01-Equivalent Grant Applications After Completion of the Kirschstein-NRSA Postdoctorate for Kirschstein-NRSA Fellows,
ORC Macro 107
With regard to the number of years to the first award, we found that Ph.D.’s and M.D./Ph.D.’s received their first R01 significantly sooner than M.D.’s.
by Degree
Note: Further details are presented in appendix table C.8.
Figure 4.17. Years to First Award for R01 and R01-Equivalent Grant Applications After Completion of the Kirschstein-NRSA Postdoctorate for Kirschstein-NRSA Trainees,
by Degree
Note: Further details are presented in appendix table C.8.
ORC Macro 108
Figure 4.18. Years to First Award for R01 and R01-Equivalent Grant Applications After Completion of the Kirschstein-NRSA Postdoctorate for Kirschstein-NRSA
Trainee/Fellows, by Degree
Note: Further details are presented in appendix table C.8.
Table 4.12. Years to First Award of an R01 or R01-Equivalent Competing Grant
Kirschstein-NRSA Group Parameter EstimateR01s1 R01 Equivalents2
Kirschstein-NRSA trainee/fellows 0.00 0.00Kirschstein-NRSA fellows 0.07 0.19Kirschstein-NRSA trainees 0.98* 0.89*
Degree GroupM.D. 0.86* 1.10*M.D./Ph.D. -0.07 0.07Ph.D. 0.00 0.00
Months of Support0–9 0.29 0.91*10–12 0.79* 1.11*13–24 0.30 0.65*25–36 0.03 0.2937+ 0.00 0.00
ContrastsFellows versus trainees Significant (fellows receive
their first R01s earlier)Significant (fellows receive their first R01 equivalents earlier)
M.D.’s versus Ph.D.’s Significant (Ph.D.’s receive their first R01s earlier)
Significant (Ph.D.’s receive their first R01 equivalents earlier)
Number of observations 8,181 9,5881 Regression data can be found in appendix table D.19.2 Regression data can be found in appendix table D.20.* Indicates statistical significance at p = .05
ORC Macro 109
4.2.4.5. Overview
This analysis focused on the success of Kirschstein-NRSA postdoctorates, identified by type of degree, in securing NIH R01 and R01-equivalent funding. We found that:
For Kirschstein-NRSA fellows:
A greater number of Ph.D.’s and M.D./Ph.D.’s secured an R01 grant. Ph.D.’s and M.D.’s obtain more grants and take less time to receive the first grant.
For Kirschstein-NRSA trainees:
A greater percentage of M.D./Ph.D.’s apply for and obtain R01 and R01-equivalent grants. They also receive a greater number of grants.
Ph.D.’s and M.D.’s have similar profiles.
For Kirschstein-NRSA trainee/fellows:
M.D./Ph.D.’s apply for and obtain more grants, although their years to first award are longer.
M.D.’s perform better than Ph.D.’s in terms of the number of individuals, although their award rate and years to first award are the same.
4.3. PUBLICATIONS
A second outcome measure is based on the extent to which an individual is represented in the peer-reviewed scientific literature. There are generally two components to publication history. The first is the number of publications that list the researcher as an author. The suggestion is that success should be reflected in the gross number of publications; however, there are a number of caveats to this success indicator. One important caveat in the biomedical sciences is that published articles are usually co-authored by several authors, and the relative contribution of the authors is not generally obvious. Thus, a researcher may be listed as an author on several publications without being a primary contributor on the research leading to those publications. In fact, in team-oriented research, this may be more likely to occur.
A second caveat relates to the quality of the published article. In addition to the quality and prestige of the journal in which the article is printed, there is the quality of the article itself. The article carries information for other researchers, and, to a degree, its importance can be measured in terms of its contribution to other research and articles. A measure of this would be the degree to which the article is cited.
Understanding that the rendering of publications data to characterize research success is a complex undertaking, we believe that, at the very least, simple publication and citation counts can provide a basis for comparing the groups. However, the caveats mentioned above must be taken into consideration when interpreting the data.
ORC Macro 110
4.3.1. Outcome Measures, Methodology, and Data Sources
The analysis in this section is largely limited by the publications and citations data extracted for consideration from Thomson ISI, a firm that compiles and maintains a database of scientific publications, their authors, and other relevant information about the publications. The extraction file was based on individuals selected from the SDR who were in the same behavioral and biomedical fields as individuals who are in the Kirschstein-NRSA postdoctoral population. The population of research doctorates selected and submitted to Thomson ISI was matched largely by individual names.62 In total, the file pertinent to this study and sent to Thomson ISI for matching comprised 2,517 research doctorates.
This section examines publications with regard to four outcome measures:
The percentage of research doctorates who have published, by number of publications and group
The mean number of publications, by group The mean elapsed time to first publication, by group The median and mean number of citations, by group
Two cohorts were examined: individuals receiving their doctorates in 1981–1982 and individuals receiving their doctorates in 1987–1988. The latter cohort has a less lengthy observation period that would presumably result in fewer publications than the earlier cohort. Inferences made by comparing the cohorts may pose some issues. For instance, it is difficult to determine whether a disparity might arise in relation to the characteristics of the cohorts, either by the potentially different environments in which they are publishing or by the differences in the career stages reached by the two cohorts. This section largely focuses on the 1981–1982 cohort, because the greater length of time allows the trends to reveal themselves more fully.
4.3.2. Results
We focus on describing differences among the six groups for each of the four specific outcome measures. We also provide a regression analysis of the number of publications.
62 More information is provided on the publications sample in appendix B of Georgine Pion. (2001). The Early Career Progress of NRSA Predoctoral Fellows and Trainees, http://grants.nih.gov/training/career_progress/.
ORC Macro 111
4.3.2.1. Percentages of Research Doctorates Who Have Published
This section examines the percentages of individuals who have published at three different levels: published at least once, more than 10 publications, and more than 20 publications. Publications are examined by type and receipt of Kirschstein-NRSA postdoctoral support and also by months of Kirschstein-NRSA support received. We look at publishing records at different thresholds to identify which groups are more likely to contain highly productive researchers.
As figures 4.19 and 4.20 show, more than 80 percent of the entire population has published at least once, with the exception of the other research doctorate group. Trainee/fellows were slightly better represented in the published population than other groups, with more than 90 percent of their members publishing at least once. Comparing the different groups reveals a distinct difference between NRSA postdoctoral recipients and other study groups.63 The data show that:
Larger percentages of Kirschstein-NRSA postdoctoral recipients produced publications at all publication levels when compared to other research doctorates. Three times more Kirschstein-NRSA postdoctoral recipients produced more than 10 publications than did other research doctorates (49, 44, and 51 percent of fellows, trainees, and trainee/fellows, respectively, versus 14 percent of other research doctorates). The same proportion was true when considering those with more than 20 publications (22, 17, and 18 percent of fellows, trainees, and trainee/fellows, respectively, versus 6 percent of other research doctorates).
Kirschstein-NRSA postdoctoral support recipients were also more successful than other non-Kirschstein-NRSA postdoctorates at all publication levels, particularly with regard to percentages of individuals who had more than 10 publications (49, 43, and 51 percent of fellows, trainees, and trainee/fellows, respectively, versus 33 percent of other postdoctorates) and more than 20 publications (22, 17, and 18 percent of fellows, trainees, and trainee/fellows, respectively, versus 11 percent of other postdoctorates).
Among Kirschstein-NRSA postdoctorates, fellows appear to be only marginally more successful than trainees; it is not notable except that it is consistent across all three publishing thresholds.
63 It should be noted that both cohorts had similar percentages of individuals who published at least once, although the 1987–1988 cohort’s percentages tend to be slightly lower, as we might expect. The 1987–1988 cohort shows a significantly more precipitous percentage drop between the published at least once and more than 10 publications levels compared with the 1981–1982 cohort, indicating again that the length of time was not sufficient to show representative results for that cohort at higher levels of publication.
ORC Macro 112
For the 1981–1982 cohort, Kirschstein-NRSA postdoctoral recipients had higher percentages publishing than other non-Kirschstein-NRSA postdoctorates and other research doctorates.
In terms of length of time receiving Kirschstein-NRSA postdoctoral support, the group with 25–36 months of support generally had the highest percentages of individuals with publications, whereas the group with 0–9 months of support had the lowest percentages, particularly in terms of highly productive individuals (i.e., those with more than 10 and those with more than 20 publications).
Trainee/fellows are more successful than all the other postdoctoral groups in terms of publishing at least once (92 percent) and having more than 10 publications (51 percent), although fellows have a higher percentage of those with more than 20 publications (22 percent of fellows versus 18 percent of trainee/fellows).
Figure 4.19. Percentage of Research Doctorates Who Have Published Between Receipt of the Doctorate and 1995, by Number of Publications and Group (1981–1982 cohort)
Sources: ISI, TFFNote: Further details are presented in appendix table C.9.
ORC Macro 113
Figure 4.20. Percentage of Research Doctorates Who Have Published Between Receipt of the Doctorate and 1995, by Number of Publications and Group (1987–1988 cohort)
Sources: ISI, TFFNote: Further details are presented in appendix table C.9.
In terms of length of time receiving Kirschstein-NRSA postdoctoral support, the group with 25–36 months of support generally had the highest percentages of individuals with publications, particularly in terms of highly productive individuals (i.e., those with more than 10 and more than 20 publications) (figure 4.21). Those with 0–9 months of support had high percentages publishing at least once, but the percentage declined sharply at the more than 10 publications level and was three to five times lower than the other categories at the more than 20 publications level. There appears to be a contrast between those with a few months of NRSA support (who publish but not frequently), and those with 2–3 years of support (who publish more frequently).
ORC Macro 114
Figure 4.21. Percentage of Research Doctorates Who Have Published Between Receipt of the Doctorate and 1995, by Months of Support Received (1981–1982 cohort)
Sources: ISI, TFF
4.3.2.2. Mean Number of Publications
While the last section focused on the percentage of individuals publishing at three specified publication thresholds, this section examines the mean number of publications.
As found in the previous analysis, all Kirschstein-NRSA postdoctoral support recipients were more successful than other postdoctorates in terms of average number of publications between receipt of doctorate and 1995 (14 for fellows, 12 for trainees, and 13 for trainee/fellows versus 9 for other non-Kirschstein-NRSA postdoctorates). Their average was also two to three times higher than that of other research doctorates (5).
Among Kirschstein-NRSA postdoctoral support recipients, fellows had the highest average number of publications with 14, compared with 12 for trainees and 13 for trainee/fellows. Unsuccessful Kirschstein-NRSA fellowship applicants had a lower average number of publications than those who received fellowships (10 versus 14, respectively). An analysis of median publication rates supports these results. Other non-Kirschstein-NRSA postdoctorates and other research doctorates had median publication numbers that were notably lower than those for Kirschstein-NRSA postdoctoral support recipients. The median number of publications for
ORC Macro 115
All Kirschstein-NRSA postdoctoral support recipients were more successful than other postdoctorates and other doctorates in terms of average number of publications.
Kirschstein-NRSA fellows appeared to have a higher average number of publications than trainees or trainee/fellows.
In terms of length of time receiving Kirschstein-NRSA support, postdoctorates with 25–36 months of support had the highest average.
Kirschstein-NRSA postdoctoral fellows was 10, twice as high as that for other non-Kirschstein-NRSA postdoctorates (5). Other research doctorates had a median of one publication.
Figure 4.22. Mean Number of Publications Between Receipt of the Doctorate and 1995, by Comparison Group (1981–1982 Cohort)
Sources: ISI, TFFNote: Further details are presented in appendix table C.10.
In terms of length of time receiving Kirschstein-NRSA support, postdoctorates with 25–36 months of support had the highest average (15), followed by those with 10–12 months and
13–24 months of support (13 and 11, respectively). Those with 25–36 months of support and those with 10–12 months of support had nearly twice the average number of publications that those in the 0–9 months group did.
ORC Macro 116
Figure 4.23. Mean Number of Publications Between Receipt of the Doctorate and 1995, by Months of Kirschstein-NRSA Postdoctoral Support Received (1981–1982 Cohort)
Sources: ISI, TFFNote: Further details are presented in appendix table C.10.
4.3.2.3. Elapsed Years From Doctorate to First Publication
This section focuses on the elapsed years to first publication after receipt of the doctorate. Kirschstein-NRSA postdoctoral support recipients, particularly trainees and trainee/fellows, tended to publish more quickly after receipt of their doctorate than other non-Kirschstein-NRSA postdoctorates and other research doctorates. Other non-Kirschstein-NRSA postdoctorates had an average elapsed time of 2.4 years and other research doctorates an average of 3 years, compared with 2.1 years for fellows, 1.5 years for trainees, and 1.9 years for trainee/fellows. Fellows and unsuccessful fellowship applicants had similar average times before first publication (2.1 and 2.2 years, respectively).
ORC Macro 117
Kirschstein-NRSA postdoctoral support recipients published more quickly after receipt of their doctorate than other non-Kirschstein-NRSA postdoctorates and other research doctorates.
Kirschstein-NRSA fellows had slightly longer elapsed times to first publication than trainees and trainee/fellows.
Although the patterns for both cohorts were similar, individuals in the 1987–1988 cohort had shorter average elapsed times to first publication than their respective categories in the 1981–1982 cohort.
In terms of length of time receiving Kirschstein-NRSA support, those with 25–36 months of support tended to have the shortest elapsed time to publication.
Figure 4.24. Mean Elapsed Years to First Publication After Receipt of the Research Doctorate, by Comparison Group (1981–1982 Cohort)
Sources: ISI, TFFNote: Further details are presented in appendix table C.11.
In terms of length of time receiving Kirschstein-NRSA support, those with 25–36 months of support tended to have the shortest elapsed time to publication (1.7 years). Those with 13–24 months and 0–9 months of support both had an average of 1.9 years, followed by those with 10–12 months of support (2.4 years). Postdoctorates with more than 36 months of support had the longest time to publication (2.6 years).
ORC Macro 118
Figure 4.25. Mean Elapsed Years to First Publication After Receipt of the Research Doctorate, by Months of Kirschstein-NRSA Postdoctoral Support Received
(1981–1982 Cohort)
Sources: ISI, TFF
4.3.2.4. Citations
This section examines the numbers of citations to publications authored by Kirschstein-NRSA postdoctoral support recipients and comparison groups. Citations are an important indicator of the publication’s influence.
Kirschstein-NRSA postdoctoral support recipients were cited more frequently on average than other non-Kirschstein-NRSA postdoctorates and other research doctorates. Kirschstein-NRSA fellows had an average of 36 citations; trainees, 28 citations; and trainee/fellows, 31 citations, compared with 18 citations for other non-Kirschstein-NRSA postdoctorates and 20 citations for other research doctorates.
As with other measures of publications, fellows had a higher average number of citations (36) than trainees (28) or trainee/fellows (31). The average number of citations for unsuccessful fellowship applicants was lower than that of fellows (29 versus 36, respectively) and slightly lower than that of trainee/fellows, although it was higher than the average for trainees.
ORC Macro 119
Kirschstein-NRSA postdoctoral support recipients were cited more frequently on average than other non-Kirschstein-NRSA postdoctorates and other research doctorates.
Kirschstein-NRSA fellows had a higher average number of citations than trainees and trainee/fellows.
In terms of length of support, Kirschstein-NRSA postdoctorates with 25–36 months of support had the highest average number of citations.
Figure 4.26. Mean Number of Citations Per Publication Published Between Receipt of the Research Doctorate and 1995, By Comparison Group (1981–1982 Cohort)
Sources: ISI, TFFNote: Further details are presented in appendix table C.12.
In terms of length of support, Kirschstein-NRSA postdoctorates with 25–36 months of support had the highest average number of citations (35), followed by those with 10–12 months of support (30), and 13–24 months of support (27). Those with 0–9 months of support and those with more than 36 months of support again were less successful (26 and 24 average citations, respectively).
Figure 4.27. Mean Number of Citations Per Publication Published Between Receipt of the Research Doctorate and 1995, By Months of Kirschstein-NRSA Postdoctoral Support
(1981–1982 Cohort)
Sources: ISI, TFF
ORC Macro 120
4.3.3. Regression Results on the Number of Publications
The preceding descriptive analyses present an overview of various publication outcomes without testing for statistical differences among the groups. The following section describes the results of a regression analysis of publication counts. Since these counts are likely to have a positive skew, we model the equation through a Poisson regression. It should be noted that cohort is embedded in the model and therefore the observation period (length of time after award) is controlled.
Table 4.13 provides the results of the regression of the publication count on the study groups and a variety of other factors. The table indicates that all NRSA postdoctorates study groups have more publications than the other non-Kirschstein-NRSA study groups.
Table 4.13. Number of Publications After Receipt of DoctorateStudy Group Parameter Estimates
NRSA fellows 0.49*NRSA trainee/fellows 0.48*NRSA trainees 0.48*Other non-Kirschstein-NRSA postdoctorates 0.29*Unsuccessful fellowship applicants 0.31*Other research doctorates 0.00
ContrastsFellows versus trainees Not significantUnsuccessful fellowship applicants versus trainees Significant (trainees publish more)Trainees versus other non-Kirschstein-NRSA postdoctorates
Significant (trainees publish more)
Number of observations 2,517* Indicates statistical significance at p = 0.05Note: Regression data can be found in appendix table D.21.
Table 4.14, which focuses only on the NRSA postdoctoral population data, basically reiterates the finding in Table 4.13: none of the NRSA groups distinguish themselves from the others. However, the data indicate that those with 0–9 months of support have poorer publication records than those with greater months of support, while those with 25–36 months have a better publication record than others.
ORC Macro 121
Table 4.14. Number of Publications After Receipt of DoctorateGroup Parameter EstimatesNRSA fellows 0.08NRSA trainee/fellows 0.00NRSA trainees -0.01
ContrastsFellows versus trainees Not significant
Months of support0–9 months -0.34*10–12 months 0.0513–24 months -0.0725–36 months 0.10*37+ months -0.00Number of observations 883* Indicates statistical significance at p = 0.05Note: Regression data can be found in appendix table D.22.
4.3.4. Summary
Conclusions from this section include the following:
Kirschstein-NRSA postdoctoral support recipients were generally more successful on all publication-related outcome measures than other non-Kirschstein-NRSA postdoctorates and other research doctorates.
There is little difference among NRSA postdoctoral study groups. Differences among fellows and trainees were marginal at best.
Higher productivity was observed for NRSA postdoctoral recipients who had 25–36 months of support and lower among those with 0–9 months of support.
Unlike the last section, where NRSA fellows outperformed NRSA trainees in securing grant funding, here we see very little difference. However, we found that there seem to be differences in elapsed time to first publications, with NRSA trainees being more likely to publish before NRSA fellows. This finding in particular may be important in that it demonstrates differences among the two groups in how they become involved in the publication process. Supporting this is the difference in the percentage of each of these two study groups that had 20 or more publications. In the 1987–1988 cohort, a greater percentage of trainees than fellows met this publication threshold. The opposite finding was noted for the 1981–1982 cohort, which may indicate that NRSA fellows start out more slowly than trainees but eventually catch up.
4.4. EMPLOYMENT
Employment is the third outcome measure that is examined in this chapter. As with the other outcome measures, the emphasis is on whether Kirschstein-NRSA fellows and trainees maintain their focus on research after completing their postdoctorates and the extent to which their careers advance in this regard. This section examines two major categories of outcome measures related to employment:
ORC Macro 122
Employment setting—Research can be conducted in a variety of settings, although academia may offer researchers the best opportunity to do so. Opportunities for conducting important research are also available in private business/industry, exemplified by firms that conduct genomic research or by research at Government facilities such as NIH. However, many research doctorates accept administrative positions in these types of organizations and do not actively engage in research.
Academic rank and tenure status—A second category of employment-related outcome measures relates to the extent to which research postdoctorates advance in their careers. For this measure, we focus on academia. In general, we would expect an orderly progression from assistant professor (and perhaps other adjunct positions) to associate professor and full professor. Each step (including achieving tenure) indicates a higher level of career attainment as a recognized researcher/educator in the field. Thus, academic rank and tenure are measures of achievement in the field and an outcome that Kirschstein-NRSA support would be expected to facilitate.
4.4.1. Outcome Measures, Methodology, and Data Sources
The outcome measures for this section are relatively simple in that they focus on the percentages of the specific population that achieved each employment outcome described above. We will examine:
The percentage of the individuals in each study group who are employed in the following settings:
Academia Business/Industry Other
The percentage of the individuals in each study group in academic settings who are in the following positions:
Assistant professor Associate professor Full professor Other
The percentage of individuals in each study group and in academic settings who are tenured Because employment status can change throughout a career, we examine these outcomes at the following three points in time:
Four years after receipt of the research doctorate—Using this timeframe, many of the individuals will be employed or at least no longer receiving postdoctoral fellowship or training support. However, about one-quarter will still be involved in postdoctoral studies.
ORC Macro 123
Eight years after receipt of the research doctorate—Almost all individuals should have begun employment, and most of those involved in research will be involved in establishing themselves as senior researchers.
Twelve years after receipt of the research doctorate—Many individuals will have established themselves as senior researchers, and many will have tenure.
The data for these analyses were obtained from the SDR, specifically for individuals who were represented in the 1995 SDR. Individuals in this survey included only those who received a research doctorate, and therefore the analysis is limited to that population. Because the SDR is a biennial longitudinal study, the 1995 SDR respondents were a mix of individuals from various research doctorate cohorts, providing information about employment outcomes at various times from 1975 to 2004. For each individual, the year that the research doctorate was awarded was identified, and responses to the SDR that was conducted 3–4 years later were tallied (the SDR is conducted every 2 years).
Table 4.15 provides the number of observations for each group. The results were weighted by the derived weighting variable supplied with the SDR. Analysis of weighted versus unweighted data showed no notable discrepancies between the two.
Table 4.15. Number of Observations in the Analytic Data File, by Study Group Study Group Number of Observations
Kirschstein-NRSA fellows 612Kirschstein-NRSA trainees 892Kirschstein-NRSA trainee/fellows 184Unsuccessful fellowship applicants 615Other non-Kirschstein-NRSA postdoctorates 2,547Other research doctorates 5,809
4.4.2. Results
As indicated above, although a majority of research doctorates were employed 4 years after they received their degrees, more than one-quarter were still involved in their postdoctoral studies. Information from the SDR also indicates a sharp decline in the percentage of individuals in postdoctoral positions 8 years after receipt of the doctorate and a corresponding increase in full-time employment.
ORC Macro 124
4.4.2.1. Employment Setting
This section focuses on the percentage of Kirschstein-NRSA postdoctoral support recipients and members of the comparison groups in three employment settings: academia,64 business/industry, and other.65 Outcomes are examined 4, 8, and 12 years after receipt of the research doctorate, while regression results are provided for 8 and 12 years after receipt of the research doctorate. Individuals who indicated that they were currently receiving postdoctoral support were eliminated from the tabulations and regression analyses.
A higher percentage of Kirschstein-NRSA fellows, trainees, and trainee/fellows were in academic employment than unsuccessful fellowship applicants, other non-Kirschstein-NRSA postdoctorates, and other research doctorates (65, 59, and 65 percent versus 46, 45, and 44 percent, respectively) 4 years after receipt of the doctorate (see table 4.16). A higher percentage of unsuccessful fellowship applicants were in business/industry than any of the other groups (22 percent).
After 8 years, the data indicate that there was little change in the percentage of Kirschstein-NRSA postdoctorates who were in academic and affiliated research settings. There was an increase among unsuccessful fellowship applicants in academic employment (an 8-percent increase for unsuccessful fellowship applicants). After 8 years, the percentage of Kirschstein-NRSA postdoctorate recipients in business and industry settings has doubled for Kirschstein-NRSA trainees and trainee/fellows and increased by half for fellows. However, the total percentage in business and industry across all groups ranged from 12 to 22 percent, still far less than the percentages in academia.
This trend for a larger proportion of the population to be employed in business/industry settings continues, as shown by the data exhibited for 12 years after receipt of the doctorate. At this point, more than one-quarter of the Kirschstein-NRSA fellows were employed in business. The relative proportion of the individuals employed in academic settings has decreased or remained stable.
Overall, Kirschstein-NRSA postdoctorate recipients differentiate themselves from other non-Kirschstein-NRSA postdoctorates and from other research doctorates in terms of being employed in academic settings and, after 12 years, in business/industry settings. 64 Includes individuals who responded that they were employed in a 1) medical school/university hospital, 2) 4-year college/university, 3) health-related school other than a medical school, or 4) university-affiliated research institute65 Includes individuals who are self-employed; employed in 1) a primary or secondary school, 2) a 2-year college, 3) Federal, State, or local government, 4) a nonacademic hospital (clinic), 5) the U.S. military, or 6) a nonprofit organization; or whose employment was classified as unknown or other in the SDR
ORC Macro 125
With regard to the percentage of each research doctorate group in various employment settings, we found that:
The percentage of individuals in business/industry increases over time in all groups.
The percentage of Kirschstein-NRSA-supported postdoctoral fellows and trainees decreases slightly between 4 and 12 years after receipt of the doctorate.
Other research doctorates, more than any other group, are employed in other types of settings.
Table 4.16. Percentage of Kirschstein-NRSA Postdoctorates and Individuals in Other Study Groups, by Employment Setting and Years After Receipt of the Doctorate
Period After Year of the Doctorate and SubgroupEmployer Setting
Business/Industry
University or Affiliated
Research FacilityOther
4 Years After DoctorateKirschstein-NRSA fellows 11 65 25Kirschstein-NRSA trainees 9 59 32Kirschstein-NRSA trainee/fellows 9 65 27Unsuccessful fellowship applicants 22 46 32Other non-Kirschstein-NRSA postdoctorates 13 45 41
Other research doctorates 9 44 488 Years After Doctorate
Kirschstein-NRSA fellows 16 65 19Kirschstein-NRSA trainees 19 57 24Kirschstein-NRSA trainee/fellows 18 61 22Unsuccessful fellowship applicants 22 54 24Other non-Kirschstein-NRSA postdoctorates 15 46 38
Other research doctorates 12 42 4712 Years After Doctorate
Kirschstein-NRSA fellows 26 57 17Kirschstein-NRSA trainees 23 54 23Kirschstein-NRSA trainee/fellows 23 56 21Unsuccessful fellowship applicants 25 53 22Other non-Kirschstein-NRSA postdoctorates 16 45 38
Other research doctorates 14 40 46Note: Further details are presented in appendix table C.13.
These results are supported by the logistic regression results presented in table 4.17. The data indicate that at 8 years after the doctorate, there are differences between Kirschstein-NRSA postdoctorate recipients and individuals in the other study groups. This is confirmed by the results at 12 years after the doctorate.
ORC Macro 126
Table 4.17. Regression Results of the Proportion of Individuals Employed at a University or Affiliated Research Facility 8 and 12 Years After Receipt of the Doctorate
GroupParameter Estimates
8 Years After Doctorate1 12 Years After Doctorate2
Kirschstein-NRSA fellows 0.80 * 0.63*Kirschstein-NRSA trainee/fellows 0.74* 0.75*Kirschstein-NRSA trainees 0.54* 0.58*Unsuccessful fellowship applicants 0.52* 0.60*Other non-Kirschstein-NRSA postdoctorates 0.18* 0.28*Other research doctorates 0.00 0.00
ContrastsFellows versus trainees Significant (fellows are
more likely to be employed at universities)
Not significant
Unsuccessful fellowship applicants versus trainees Not significant Not significantTrainees versus other non-Kirschstein-NRSA postdoctorates
Significant (trainees are more likely to be employed at universities)
Significant (trainees are more likely to be employed at universities)
Number of observations 7,660 7,2901 Regression data can be found in appendix table D.23.2 Regression data can be found in appendix table D.24.* Indicates statistical significance at p = 0.05
4.4.3. Academic Rank and Tenure Status
The analysis in the previous section was affected in part by the success of individuals in an academic setting in progressing through the ranks and obtaining tenure. Individuals having difficulty doing so may look for employment in business/industry or as administrators in government or the nonprofit sector. This section examines outcomes related to academic rank (assistant professor, associate professor, professor, and other) and tenure status. The analysis focuses on only those individuals in academic and affiliated research settings (as defined in section 4.4.2.1).66
Four years after receipt of the doctorate, the percentages of research doctorates with assistant professor positions ranged from 56 percent for Kirschstein-NRSA trainees to 75 percent for other non-Kirschstein-NRSA postdoctorates (see table 4.18). However, there were higher percentages of individuals in the other research doctorates study group in associate professor and full professor positions (19 percent overall), which means that a higher percentage of this group held
66 The denominator in calculating the percentage excludes individuals who have been passed over for promotion and exit academia or associated research settings by the time limit. Thus, the percentages reflect the number who have achieved the stated outcome relative to those who may still be in contention for the stated outcome.
ORC Macro 127
With regard to academic rank and tenure status outcomes, we found that:
After 4 years, very few Kirschstein-NRSA postdoctorates achieve associate or full professorship or tenure, when compared with other research doctorates and other non-Kirschstein-NRSA postdoctorates.
Although a gap continues up until 12 years, it decreases in size.
Kirschstein-NRSA fellows tend to outperform Kirschstein-NRSA trainees in terms of the percentage reaching an associate or full professorship and receiving tenure after 12 years.
assistant, associate, and full professor positions, rather than other academic positions. These individuals seem to have progressed more quickly in their academic rank standings than the Kirschstein-NRSA groups.
After 8 years, the percentage of research doctorates who were in associate and full professor positions increased in all study groups. This increase, in percentage point terms, was particularly notable among the non-Kirschstein-NRSA study groups. Other research doctorates continued to have the highest percentages of individuals at the associate professor and full professor levels (53 and 10 percent, respectively).
However, Kirschstein-NRSA fellows were more represented than the other study groups at the associate professor rank 12 years after receipt of the doctorate. They still did not equal the proportions of other research doctorates at the full professor position (5 and 31 percent, respectively), but the data suggest that because they enter the academic job market later, it might take them longer to reach these senior levels. The percentages of individuals employed as assistant professors decreased and as associate professors increased for all groups between 8 and 12 years after receipt of the doctorate.
Table 4.19 presents logistic regression results of the proportion of individuals attaining associate or full professor positions 8 and 12 years after receipt of the doctorate. The regressions show that, generally, after 8 years the Kirschstein-NRSA groups were less likely to obtain an associate or full professorship than non-Kirschstein-NRSA postdoctorates, and after 12 years, Kirschstein-NRSA fellows were more likely than Kirschstein-NRSA trainees to obtain an associate or full professorship. Kirschstein-NRSA trainees did not perform as well as other Kirschstein-NRSA postdoctorates in advancing in academic rank. Although there seems to be a gap between the academic rank of those with Kirschstein-NRSA postdoctoral support and the non-Kirschstein-NRSA study groups, Kirschstein-NRSA fellows seem to do better than trainees at decreasing this gap over time. It should be noted that these comparisons excluded individuals who left academia at some point and thus examined only those individuals who were successful or still have the opportunity to be successful with regard to this outcome.
ORC Macro 128
Table 4.18. Percentage of Kirschstein-NRSA Postdoctorates and Individuals in Other Study Groups, by Academic Rank and Years After Receipt of the Doctorate
Period After Year of Doctorate and SubgroupAcademic Rank
AssistantProfessor
AssociateProfessor
FullProfessor Other
4 Years After DoctorateKirschstein-NRSA fellows 69 3 0 28Kirschstein-NRSA trainees 66 3 0 31Kirschstein-NRSA trainee/fellows 56 6 0 38Unsuccessful fellowship applicants 69 2 1 29Other non-Kirschstein-NRSA postdoctorates 75 5 0 20
Other research doctorates 67 16 3 148 Years After Doctorate
Kirschstein-NRSA fellows 72 16 0 12Kirschstein-NRSA trainees 66 17 1 16Kirschstein-NRSA trainee/fellows 77 12 0 12Unsuccessful fellowship applicants 70 14 3 14Other non-Kirschstein-NRSA postdoctorates 51 34 1 13
Other research doctorates 29 53 10 912 Years After Doctorate
Kirschstein-NRSA fellows 28 63 5 4Kirschstein-NRSA trainees 36 47 6 11Kirschstein-NRSA trainee/fellows 37 50 7 7Unsuccessful fellowship applicants 28 49 6 17Other non-Kirschstein-NRSA postdoctorates 18 54 18 10
Other research doctorates 9 50 31 10Note: Further details are presented in appendix table C.14.
ORC Macro 129
Table 4.19. Regression Results of the Proportion of Individuals Employed as an Associate or Full Professor 8 and 12 Years After Receipt of the Doctorate
GroupParameter Estimates
8 Years After Doctorate1 12 Years After Doctorate2
Kirschstein-NRSA fellows -1.22* -0.47*Kirschstein-NRSA trainees -1.07* -0.96*Kirschstein-NRSA trainee/fellows -1.52* -0.70*Unsuccessful fellowship applicants -1.13* -0.65*Other non-Kirschstein-NRSA postdoctorates -0.60* -0.38*Other research doctorates 0.00 0.00
ContrastsFellows versus trainees Not significant Significant (fellows are more
likely to attain an associate or full professorship)
Unsuccessful fellowship applicants versus trainees Not significant Not significantTrainees versus other non-Kirschstein-NRSA postdoctorates
Significant (trainees are less likely to attain an associate or full professorship)
Significant (trainees are less likely to attain an associate or full professorship)
Number of observations 2,929 2,5981 Regression data can be found in appendix table D.25.2 Regression data can be found in appendix table D.26.* Indicates statistical significance at p = 0.05
Employment in academic positions usually involves the receipt of tenure. We present the results with regard to tenure in tables 4.20 and 4.21. The results are similar to those discussed in reference to academic rank. As we would expect, at the 4-year period tenure was rare. Only 19 percent of the other research doctorates study group and 8 percent of the other non-Kirschstein-NRSA postdoctorates study group had tenure at this point in time. This compares with almost none in the Kirschstein-NRSA postdoctorate groups.
After 8 years, these percentages increase dramatically, but the ordering remains much the same. Thus, about 65 percent of the other research doctorates study group and 40 percent of the other non-Kirschstein-NRSA postdoctorates study group had received tenure. No other study group had more than 20 percent of its population tenured.
This same pattern is apparent after 12 years, where the percentage of tenured individuals was highest in the other non-Kirschstein-NRSA postdoctorates study group (71 percent) and other research doctorates study group (79 percent). However, Kirschstein-NRSA fellows, trainees, and trainee/fellows all experienced a large increase in the percentages achieving tenure; all of these groups had increases approaching 40 percentage points. Kirschstein-NRSA fellows and trainees still had percentages that were lower than other research doctorates and other non-Kirschstein-NRSA postdoctorates, but the gap between the groups was narrower.
The results from the logistic regression confirm that Kirschstein-NRSA postdoctorates are less likely to obtain tenure at both the 8- and 12-year periods but that there seems to a “catch up” factor when compared with other research doctorates.
ORC Macro 130
Table 4.20. Percentage of Kirschstein-NRSA Postdoctorates and Individuals in Other Study Groups, by Tenure Status and Years After Receipt of the Doctorate
SubgroupPercentage Who Were Tenured
4 Years After Doctorate
8 Years After Doctorate
12 Years After Doctorate
Kirschstein-NRSA fellows 1 13 55Kirschstein-NRSA trainees 3 18 49Kirschstein-NRSA trainee/fellows 5 7 55Unsuccessful fellowship applicants 0 16 54Other non-Kirschstein-NRSA postdoctorates
8 40 71
Other research doctorates 19 65 79Note: Further details are presented in appendix table C.14.
Table 4.21. Regression Results of the Proportion of Individuals Tenured 8 and 12 Years After Receipt of the Doctorate
GroupParameter Estimates
8 Years After Doctorate1 12 Years After Doctorate2
Kirschstein-NRSA trainee/fellows -2.14* -0.12Kirschstein-NRSA fellows -1.38* -0.35Kirschstein-NRSA trainees -1.18* -0.57*Unsuccessful fellowship applicants -1.12* -0.17Other non-Kirschstein-NRSA postdoctorates -0.57* -0.10Other research doctorates 0.00 0.00
ContrastsFellows versus trainees Not significant Not significantUnsuccessful fellowship applicants versus trainees Not significant Not significantTrainees versus other non-Kirschstein-NRSA postdoctorates
Significant (trainees are less likely to achieve tenure)
Significant (trainees are less likely to achieve tenure)
Number of observations 2,576 2,5101 Regression data can be found in appendix table D.27.2 Regression data can be found in appendix table D.28.* Indicates statistical significance at p = 0.05
4.4.4. Summary
The analysis of employment setting, academic rank, and tenure indicates that after 4 years, Kirschstein-NRSA postdoctorates are not as advanced in their careers as other research doctorates or other non-Kirschstein-NRSA postdoctorates. The differences between the non-Kirschstein-NRSA and Kirschstein-NRSA study groups decrease in subsequent periods, indicating that Kirschstein-NRSA postdoctorates have accelerated career paths. After 12 years, outcomes for Kirschstein-NRSA postdoctorates and other non-Kirschstein-NRSA postdoctorates are similar. It is important to note that the relative proportions of postdoctorates entering business/industry increase over time, which is probably the result of tenure and advancement outcomes in academic settings. It should also be noted that the apparent lag in tenure and position may reflect the type of academic setting in which employment takes place. For example, other research doctorates may be more likely to be employed in teaching rather than research settings.
ORC Macro 131
4.5. SUMMARY AND CONCLUSIONS
Our analysis in this chapter focused on outcomes related to research productivity and measures of research status. A key element of this analysis is the supposition that Kirschstein-NRSA fellowship and traineeship support facilitates these outcomes. We examined three sets of outcomes: success in obtaining research funding from NIH, publication rates, and employment.
With regard to success in obtaining research funding from NIH, we found the following (see table 4.22):
On most measures related to grant outcomes, Kirschstein-NRSA fellows performed better than most groups.
Kirschstein-NRSA trainees, however, performed at or below the level of unsuccessful fellowship applicants and other non-Kirschstein-NRSA postdoctorates.
Kirschstein-NRSA trainees who also received fellowship support performed similar to fellows on most measures.
Some of the measures indicate that receiving Kirschstein-NRSA postdoctoral support prepares individuals to apply for and be awarded grants other than R01s.
Some of the measures, such as years to first R01 award, indicate that Kirschstein-NRSA postdoctoral recipients experience a lag in obtaining R01 support when compared with other research doctorates who presumably become employed shortly after the completion of their doctorates.
With regard to comparisons between Ph.D.’s, M.D.’s, and those with M.D./Ph.D.’s, we found that there were differences in success rates between the three groups (see table 4.23).
For Kirschstein-NRSA fellows, Ph.D.’s and MD/Ph.D.’s had higher award rates and fewer years to first award than those with only an M.D.
For Kirschstein-NRSA trainees, M.D./Ph.D.’s differentiated themselves from the other two degree groups.
Overall, the type of training seems to be a more consistent factor in explaining differences in the outcomes, although there is evidence of interaction effects between type of support and degree.
With regard to publication rates, we found the following (see table 4.24):
Kirschstein-NRSA postdoctoral support recipients were generally more successful on all outcome measures than other postdoctorates and other research doctorates.
There is little difference among Kirschstein-NRSA postdoctoral study groups. Differences between fellows and trainees were marginal.
Higher productivity was observed for Kirschstein-NRSA postdoctoral recipients who had 25–36 months of postdoctoral support; lower productivity was observed for those with 0–9 months of support.
ORC Macro 132
With regard to employment, we found the following (see table 4.25):
After 4 years, Kirschstein-NRSA postdoctorates are not as advanced in their careers as other research doctorates or other non-Kirschstein-NRSA postdoctorates.
The differences between the non-Kirschstein-NRSA and Kirschstein-NRSA study groups grow smaller in subsequent time periods, thus indicating that Kirschstein-NRSA postdoctorates have accelerated career paths.
After 12 years, the differences between Kirschstein-NRSA postdoctorates and other non-Kirschstein-NRSA postdoctorates have decreased. It is important to note that the relative proportions of postdoctorates entering business and industry increase over time, which is most likely a combined result of tenure and advancement outcomes in academic settings.
ORC Macro 133
Table 4.22. Summary of Grant Outcomes for Research Doctorates
OutcomesKirschstein-
NRSA Fellows (F)
Kirschstein-NRSA
Trainees (T)
Kirschstein-NRSA
Trainee/Fellows
(TF)
Unsuccessful Fellowship Applicants
(UFA)
Other Postdoctorates
(OP)
Other Research Doctorate
(ORD)
Groups Indicating Research-Oriented Outcomes*
UFA Vs T F Vs T T VsOP
Percentage applying
R01 45% 30% 46% 30% 20% 9% NS F T
R01 equivalents 53% 38% 57% 35% 23% 10% T F T
Percentage of those applying receiving at least one award
R01 61% 47% 57% 54% 49% 36% UFA F NS
R01 equivalents 67% 53% 65% 57% 51% 38% UFA F NS
Percentage with more than one grant award
R01 48% 37% 44% 46% 49% 32% Not Estimated
Not Estimated
Not Estimated
R01 equivalents Not Estimated Not
EstimatedNot
EstimatedNot
Estimated
Number of awards
R01 Not Estimated UFA F NSR01 equivalents Not Estimated UFA F NS
Award rateR01 34% 25% 30% 29% 28% 19% UFA F OPR01 equivalents 36% 27% 32% 30% 28% 19% UFA F NS
Average years to first award
R01 7.8 8.9 8.4 8.0 7.9 7.4 UFA F OP
R01 equivalents 6.6 7.2 6.8 7.0 6.9 6.8 NS F NS
* Each column represents the comparison, and each cell contains the group designation of those who rated higher with regard to the outcome.NS = not significantR01 equivalents include R01, R23, and R39 grants.
ORC Macro 134
Table 4.23. Summary of Grant Outcomes for All Doctorates
OutcomesKirschstein-NRSA Fellows (F) Kirschstein-NRSA
Trainees (T)Kirschstein-NRSA
Trainee/Fellows(TF)
Groups Indicating Research-Oriented
Outcomes*
F vs. T Ph.D. vs.M.D.
Ph.D.’s M.D.’s M.D./Ph.D. Ph.D.’s M.D.’s M.D./Ph.D. Ph.D.’s M.D.’s M.D./Ph.D.
Percentage applying
R01 49% 51% 48% 33% 33% 42% 52% 54% 60% F M.D.
R01 equivalents 54 60 55 39 40 49% 59 64 70 F M.D.
Percentage of those applying receiving at least one award
R01 63 55 63 50 50 57 62 57 60 F NS
R01 equivalents 68 58 70 55 54 61 68 61 66 F NS
Percentage with more than one grant award
R01 58 50 55 47 46 52 54 51 59 F Ph.D.
R01 equivalents
Number of awards
R01 F NS
R01 equivalents F Ph.D.
Award rate R01 31 25 30 24 25 28 28 27 29 F NSR01 equivalents 33 26 32 26 26 29 30 28 31 F NS
Average years to first award
R01 5.3 6.8 5.5 6.5 7.6 6.4 5.3 6.2 5.4
R01 equivalents 4.1 5.2 4.1 4.8 6.1 5.0 3.7 4.6 3.5
* Each column represents the comparison, and each cell contains the group designation of those who rated higher with regard to the outcome.NS = not significantR01 equivalents include R01, R23, and R39 grants.
ORC Macro 135
Table 4.24. Summary of Publications Outcomes to 1995 (1981–82 Cohort)
Publications (1981–82 Research Doctorate Cohort)
Kirschstein-NRSA
Fellows
Kirschstein-NRSA
TraineesKirschstein-NRSA
Trainee/FellowsUnsuccessful
Fellowship Applicants
Other Postdoctorates
Other Research
DoctoratesPercentage who published at least one article 86% 84% 92% 83% 82% 64%
Average number of publications 14.4 11.9 12.6 10.2 9.4 4.9
Average years to first publication 2.1 1.9 1.5 2.2 2.4 3.0
Average number of citations 36.1 27.8 30.6 29.2 18.0 19.5
ORC Macro 136
Table 4.25. Summary of Employment Outcomes
Employment Outcomes (12 Years After Award of the
Research Doctorate)
Kirschstein-NRSA
Fellows (F)
Kirschstein-NRSA
Trainees (T)
Kirschstein-NRSA
Trainee/Fellows
(TF)
Unsuccessful Fellowship Applicants
(UFA)
Other Postdoctorates
(OP)
Other Research
Doctorates(ORD)
Groups Indicating Research-Oriented Outcomes*
UFA Vs T F Vs T T VsOP
Employment Setting
Percent employed in academe
57% 54% 56% 53% 45% 40% NS NS T
Percent employed in business/industry
26% 23% 23% 25% 16% 14% Not Estimated
Not Estimated
Not Estimated
Percentage reaching associate or full professorship rank
Associate professorship 63% 47% 50% 49% 54% 50% Not
EstimatedNot
EstimatedNot
Estimated
Full professorship 5% 6% 7% 6% 18% 31% Not
EstimatedNot
EstimatedNot
EstimatedAssociate and full professorship 68% 53% 57% 55% 72% 81% NS F NS
Percentage receiving tenure of those eligible 55% 49% 55% 54% 71% 79% NS NS OP
* Each column represents the comparison, and each cell contains the group designation of those who rated higher with regard to the outcome.NS = not significant
ORC Macro 137
4.5.1. Conclusions
These results indicate differences between the Kirschstein-NRSA postdoctoral groups and the other research doctoral study groups defined for this study. However, these results largely reflect the achievements of Kirschstein-NRSA fellows and those who held both traineeships and fellowships, but not those appointed only to traineeships. Except for publication rates, this latter group in most regards is equivalent to the other non-Kirschstein-NRSA study groups. It should be noted that the activities that fellows and trainees pursue during their postdoctoral studies may differ. Fellows have their own research focus, as expressed in their Kirschstein-NRSA fellowship application; trainees generally work closely on an existing project. To better understand these differences, more information is needed on the range of activities pursued in each support category. Such information will foster an understanding of early- and late-career outcome differences.
A second conclusion indicates that Kirschstein-NRSA postdoctorates experience a delay in their careers. When compared with other research doctorates, Kirschstein-NRSA fellows in particular are slower to reach tenure and associate and full professor positions, and are slower to secure NIH funding. It should be noted that these comparisons were made not against the total subgroup of other research doctorates, but against those who obtained academic positions immediately, without the need for postdoctoral study.
ORC Macro 138
CHAPTER 4 ADDENDUM: ANALYSIS OF NIH GRANT APPLICATION AND AWARD RATES FROM 1990 TO 2002
Chapter 4 focused on outcomes for individuals whose last year of postdoctoral support was 1992 and who received a research doctorate by 1992. One such outcome explored was the success of these individuals in obtaining NIH research funding. The analyses, however, do not provide results for individuals who have more recently completed research doctorates or their Kirschstein-NRSA postdoctoral fellowships or traineeships. To determine whether the results in chapter 4 can apply to more recent Kirschstein-NRSA fellow and trainee populations, this section provides an analysis of NIH research grant outcomes for those who received their research doctorates between 1990 and 2002. We focus on the differences among various Kirschstein-NRSA-supported fellows and trainees and those who had postdoctoral plans but were not awarded or appointed to a Kirschstein-NRSA fellowship or traineeship.
Although it may be interesting to also examine recent trends for those with dual M.D./Ph.D.’s and those with M.D.’s only, we have restricted this analysis to Ph.D.’s only for the following reasons:
For M.D./Ph.D.’s, the total number of individuals examined averaged 161 per cohort (determined by year of Ph.D.) with an average of 81, or more than half, receiving Kirschstein-NRSA postdoctoral training support. The small overall numbers of these individuals who apply for and receive awards, the proportionally large number who are trainees, and the declining trend in application submissions all suggest that the number of observations do not support a robust analysis of this group. Even if aggregated over the entire 1990–2002 period, the declining numbers in more recent years suggest that the data may not be comparable with results from earlier years, in which there were a larger number of observations.
For the analysis of the M.D. population, a different issue arises. The number in each cohort may provide a basis for an acceptable exploration of the data, although our confidence in the data decreases when considering data from recent years due to the drop in the number of observations. However, more than 90 percent of the M.D. population received Kirschstein-NRSA traineeship support during this period, leaving very few observations in the other groups. Thus, whereas we could probably find evidence for trends in the trainee group during this period, comparisons with Kirschstein-NRSA fellows or Kirschstein-NRSA trainee/fellows would be very limited.
Therefore, our analysis focuses on NIH grant application rates for individuals receiving their Ph.D.’s between 1990 and 2002. There are five comparison groups:
Kirschstein-NRSA fellows—Includes those who received only fellowship support Kirschstein-NRSA trainees—Includes those who received only traineeship support Kirschstein-NRSA trainee/fellows—Includes those who received both traineeship and
fellowship support Unsuccessful Kirschstein-NRSA fellowship applicants—Includes those who applied for a
fellowship but did not receive one and who were not appointed to a traineeship
ORC Macro 139
Other research doctorates—Includes research doctorates who indicated that they had plans to pursue postdoctoral training as indicated by their response recorded in the SED and who were not in any of the above groups
Except for the other research doctorates group, these definitions are consistent with those used in the primary analysis in previous chapters. It should be noted that the total number of research doctorates varies but remains relatively constant throughout the analysis period (see tables 4A.1–4A.5). It should also be noted that our previous analysis (described in chapter 4) indicates an average lag of 5–6 years from receipt of the doctorate to first NIH research grant award. This would mean that the most recent cohorts would contain very few grant applicants, and even fewer would be expected to receive an award. Thus, in recent years there is an overall declining trend in application rates, and the instability in award rates is a result of the decline in the number of applications, which serve as the denominator for calculating award rates.
Table 4A.1. Number of Research Doctorates, by Study Group and Ph.D. Cohort
Study Group Year Received Ph.D.1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Kirschstein-NRSA fellows 288 331 317 363 386 423 344 336 313 301 272 245 230
Kirschstein-NRSA trainees 567 655 689 732 714 757 687 697 593 613 627 585 565
Kirschstein-NRSA trainee/fellows 135 139 152 169 173 172 170 143 142 131 150 81 90
Unsuccessful fellowship applicants 295 322 315 370 406 355 376 268 233 202 201 144 50
Other research doctorates 2,285 2,381 2,554 2,683 2,775 2,883 3,196 2,905 3,364 3,459 3,565 3,684 3,763
Total 3,570 3,828 4,027 4,317 4,454 4,590 4,773 4,349 4,645 4,706 4,815 4,739 4,698
Table 4A.2. Number of Research Doctorates Who Applied for NIH RPG Awards, by Study Group and Ph.D. Cohort
Study Group Year Received Ph.D.1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Kirschstein-NRSA fellows 143 150 133 170 152 148 112 101 85 51 24 13 6
Kirschstein-NRSA trainees 224 273 255 246 210 220 179 160 82 80 60 29 9
Kirschstein-NRSA trainee/fellows 72 78 76 72 81 61 69 40 25 24 12 4 1
Unsuccessful fellowship applicants 125 109 104 117 113 98 84 40 30 18 17 7 2
Other research doctorates 367 329 357 354 296 266 280 228 206 136 90 73 44
Total 931 939 925 959 852 793 724 569 428 309 203 126 62
ORC Macro 140
Table 4A.3. Number of Research Doctorates Who Received NIH RPG Awards, by Study Group and Ph.D. Cohort
Study Group Year Received Ph.D.1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Kirschstein-NRSA fellows 95 102 102 112 104 79 65 53 36 30 9 3 2
Kirschstein-NRSA trainees 135 156 141 131 118 116 93 75 47 32 24 12 4
Kirschstein-NRSA trainee/fellows 49 54 54 45 49 36 42 16 12 12 3 3 0
Unsuccessful fellowship applicants 80 70 69 69 64 57 45 19 7 6 9 3 0
Other research doctorates 227 202 195 189 171 144 140 123 95 61 33 40 20
Total 586 584 561 546 506 432 385 286 197 141 78 61 26
Table 4A.4. Number of Research Doctorates Who Applied for R01-Equivalent Awards, by Study Group and Ph.D. Cohort
Study Group Year Received Ph.D.1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Kirschstein-NRSA fellows 129 133 112 145 128 126 80 68 57 31 11 8 3
Kirschstein-NRSA trainees 199 219 209 194 169 166 124 115 51 40 25 9 1
Kirschstein-NRSA trainee/fellows 65 71 74 58 65 43 53 32 19 13 4 1 0
Unsuccessful fellowship applicants 111 96 88 97 89 79 68 31 22 11 10 3 0
Other research doctorates 317 283 292 279 230 201 214 160 116 81 36 25 11
Total 821 802 775 773 681 615 539 406 265 176 86 46 15
Table 4A.5. Number of Research Doctorates Who Received R01-Equivalent Awards, by Study Group and Ph.D. Cohort
Study Group Year Received Ph.D.1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Kirschstein-NRSA fellows 81 92 80 93 84 58 44 31 23 16 3 2 1
Kirschstein-NRSA trainees 105 116 111 92 85 77 59 43 24 10 8 4 0
Kirschstein-NRSA trainee/fellows 41 46 51 39 36 24 29 10 8 4 1 1 0
Unsuccessful fellowship applicants 68 61 57 59 52 46 31 11 5 2 6 0 0
Other research doctorates 195 168 154 153 119 108 103 78 45 29 14 12 1
Total 490 483 453 436 376 313 266 173 105 61 32 19 2
ORC Macro 141
The results of our analysis are summarized below.
NIH RPG award and R01-equivalent grant rate data show similar patterns (see figures 4A.1 and 4A.2). The data show that:
There is a constant decline in the number of grant applications submitted across all groups.
There is a distinct positive difference favoring those with Kirschstein-NRSA support over other research doctorates who had plans to pursue postdoctoral support.
Kirschstein-NRSA fellows and trainee/fellows have higher application rates than the other groups.
For recent cohorts, those with Kirschstein-NRSA fellowships apply at a greater rate than other groups (an inverse of the relationship that this group has with trainee/fellows).
The difference between Kirschstein-NRSA trainees and unsuccessful fellowship applicants was mixed. In some years, a higher proportion of trainees applied; in other years, a higher proportion of unsuccessful fellowship applicants applied.
Figure 4A.1. Percentage of Ph.D.’s Who Applied for NIH RPG Awards, by Ph.D. Year and Kirschstein-NRSA Postdoctoral Status
ORC Macro 142
Figure 4A.2. Percentage of Ph.D.’s Who Applied for R01-Equivalent Awards, by Ph.D. Year and Kirschstein-NRSA Postdoctoral Status
There are similar patterns for those who received RPG awards and R01-equivalent grants (see figures 4A.3 and 4A.4):
Trainees, at least early during the period, show a lower award rate than other groups, even those who had expressed plans to pursue a postdoctorate but did not apply for or receive Kirschstein-NRSA postdoctoral funding.
Kirschstein-NRSA fellows, trainee/fellows, and unsuccessful fellowship applicants show higher levels of success in obtaining NIH grant funding.
After 1995 the patterns become more difficult to determine due to the smaller number of individuals in the denominator (i.e., grant applicants).
ORC Macro 143
Figure 4A.3. Percentage of NIH RPG Applicants Who Received Awards, by Ph.D. Year and Kirschstein-NRSA Postdoctoral Status
Figure 4A.4. Percentage of R01-Equivalent Grant Applicants Who Received Awards, by Ph.D. Year and Kirschstein-NRSA Postdoctoral Status
CONCLUSIONS
In general, the results of this analysis support the results described in chapter 4. As more recent cohorts are considered, however, the results are less conclusive because of the small numbers of individuals involved in grant application activity. The major conclusion is more of a methodological one than a substantive one in that to be able to fully compare the success of Kirschstein-NRSA fellows and trainees with that of other research doctorates, intermediate and long-term outcomes must be examined. Examining success in receiving NIH research grant
ORC Macro 144
funding in the short term allows us to identify the few individuals who have the professional maturity to succeed in obtaining NIH research grant funding, but this will exclude the majority of individuals who, on the average, strive for several years before achieving success.
ORC Macro 145
5. SUMMARY AND CONCLUSIONS
Since its implementation in fiscal year (FY) 1975, the Kirschstein-NRSA program has been focused on supporting predoctoral and postdoctoral training in a wide range of health-related disciplines where there is a need for personnel.67 Two kinds of awards have been made: 1) institutional training grants to public and nonprofit private institutions for multidisciplinary training programs or targeted training programs in specific research areas, and 2) fellowships to individuals for supervised study with a senior scientist at an institution of the fellow’s choice. Both awards are conferred after competitive peer review. When NIH first awarded research training support in FY 1975, awards were made to 2,202 postdoctoral trainees and fellows; by FY 2002, 6,281 individuals received postdoctorate-level support.68
Kirschstein-NRSA programs that provide postdoctoral support to research and professional doctorates have entered their fourth decade. Individuals making use of the training and fellowship programs during the early years of the program have had the opportunity to enter and prosper in research careers in academia or business. Since the Kirschstein-NRSA programs are an investment in human capital for advancing scientific research, it is useful to assess the scientific contributions associated with individuals who received support. This study examines three types of outcomes related to potential scientific contributions:
Success in obtaining NIH research grant support Success in publishing in peer-reviewed journals Success in obtaining and remaining in employment
Concentrating on these outcomes provides a perspective on how well individuals supported by Kirschstein-NRSA have performed throughout their careers compared with other equivalent groups of doctorates.
5.1. STUDY OUTCOMES
5.1.1. Success in Obtaining NIH Research Grant Support
NIH funds a proportionally large percentage of U.S. biomedical research through various grant mechanisms. The most important mechanism is the R01 grant, which is competed through a peer-review process. An R01 grant implies that the scientific community has a high level of confidence that the project supported by the grant will advance biomedical and behavioral knowledge. These grants are generally provided to principal investigators who have a track record in research. To assist new investigators, the New Investigator Research Award (NIRA) (R23) and First Independent Research Support and Transition (FIRST) Award (R29) programs were instituted. Although they have since been discontinued,69 they provided an alternative to 67 The National Academy of Sciences, through its Committee on National Needs for Biomedical and Behavioral Scientists, has periodically issued reports specifying the number and type of trainees who should be funded under the Kirschstein-NRSA program.68 NIH Trainee and Fellow File (2004)69 The NIRA (R23) award mechanism began in 1977; it was superseded by the FIRST (R29) award mechanism, which was discontinued in 1998.
ORC Macro 146
R01 grants while they were active. We therefore consider the R23 and R29 grant mechanisms to be R01 equivalents. Our analysis focuses on these two categories of grants (R01 and R01 equivalents) because they constitute the primary funding vehicles for NIH. In assessing success in obtaining NIH R01 and R01-equivalent support, we use the following measures:
Percentage of Doctorates Applying for These Grants provides an indication of the interest in pursuing research in the population of doctorates.
Percentage of Doctorates Being Awarded At Least One Grant provides an indication of the prevalence of success in obtaining research funding.
Number of Grants Awarded to Individuals provides an indication of the overall success of individuals throughout their careers.
Average Individual Award Rate provides an indication of individuals’ overall ability to obtain funding throughout their careers.
Time Before First Award provides information on the wait before individuals achieve the status of awarded principal investigator—an important step in their career development.
5.1.2. Success in Publishing in Peer-Reviewed Journals
Publications provide evidence of scholarly achievement, especially if the focus is on publication in peer-reviewed journals. The number of publications and their quality are used to gauge promotions in academic institutions and whether tenure is awarded. One measure, the number of publications, demonstrates the involvement of researchers in scientific discourse. A second measure, the number of citations, relates to the attention the publication receives from the scientific community. Frequently cited articles are thought to have more of an impact. The following measures were examined:
Percentage of Study Population Who Are Published—This measure presents the prevalence of individuals authoring at least one peer-reviewed publication. In other words, it measures the extent to which the population contributes to scientific discourse in a field of study.
Number of Publications—For those who published, this measure presents a indicator of the extent of participation by individuals in scientific discourse in a field of study.
Years to First Publication—This measures the time it takes to produce the first publications for those who have published after receipt of the Ph.D.
Average Number of Citations—This number reflects the influence of the articles published by individuals in the various study groups.
5.1.3. Success in Obtaining and Remaining in Research-Oriented Employment
An expectation for postdoctoral researchers is that they will continue with research. Perhaps the best positions that will allow individuals to do this are in academia or selected research institutions. Although positions in the business and industry sector offer interesting opportunities, these opportunities are not as unencumbered as those in academia. We examined the following measures with regard to employment:
ORC Macro 147
Employment and Employment in Academia—At the very least, it would be of interest to examine the employment status of individuals at different stages in their careers. Employment in academia would provide an indicator of the individual’s pursuit of research or teaching positions.
Academic Position and Tenure—Career advancement in academia is usually observed through promotions from junior positions (e.g., adjunct or assistant professor) to senior positions (e.g., associate or full professor) and by the award of tenure.
5.2. STUDY GROUPS
In examining outcomes, the most obvious comparison is between Kirschstein-NRSA postdoctoral researchers and individuals not receiving Kirschstein-NRSA support. However, this comparison is problematic in that the former group contains individuals with different kinds of support, primarily Kirschstein-NRSA traineeships and fellowships, and different lengths of support. Differences in Kirschstein-NRSA support may have an effect on outcomes. With regard to those without Kirschstein-NRSA support, we may again find a wide range of outcomes: For example, in terms of employment outcomes, individuals may find another type of postdoctoral support to continue with their studies, they may accept teaching or research jobs, or they may decide to focus on administrative or other nonresearch positions. This may also be manifested in terms of grants and publications.
In chapter 3, we examined the nature of the Kirschstein-NRSA experience and the kinds of information about that experience that are important for describing useful groups for the analysis. The analysis showed that Kirschstein-NRSA is not a unitary intervention, i.e., that the training provided is relatively diverse both in kind and length. The analysis in that chapter revealed that both type (a traineeship appointment or an individual fellowship award) and length of Kirschstein-NRSA support are important.
Using results from that analysis, we identified the following six study groups:
Kirschstein-NRSA Fellows—This group includes doctorates who received a Kirschstein-NRSA fellowship through an F32 or F35 grant mechanism. It includes individuals who applied more than once prior to being awarded a fellowship, but excludes Kirschstein-NRSA fellows who were appointed to a Kirschstein-NRSA traineeship.
Kirschstein-NRSA Trainees—This group includes doctorates who were appointed to a Kirschstein-NRSA traineeship by an institution with a T32 (training) grant. It excludes individuals who received both a traineeship and a fellowship during their postdoctoral tenure.
Kirschstein-NRSA Trainee/Fellows—This group includes doctorates who were supported
by both a fellowship and traineeship during their postdoctoral tenure. Since it may have characteristics of both trainees and fellows, we believed that separating this group from the other Kirschstein-NRSA-supported groups might eliminate any ambiguities concerning comparisons between fellows and trainees.
ORC Macro 148
Unsuccessful Kirschstein-NRSA Fellowship Applicants—This group includes doctorates who applied for but did not receive a Kirschstein-NRSA fellowship. It includes individuals who may have received formal postdoctoral funding from other sources.
Other Non-Kirschstein-NRSA Postdoctorates—This group includes research doctorates who were appointed to a postdoctoral position supported by non-Kirschstein-NRSA sources such as NSF. It excludes individuals who applied for a Kirschstein-NRSA fellowship.
Other Research Doctorates—This group includes research doctorates who had no indicated postdoctoral activity, including the submission of an application for a Kirschstein-NRSA fellowship.
Using these study groups, we can generate a wide variety of comparisons to investigate groups that are hypothesized to be roughly equivalent (e.g., Kirschstein-NRSA fellows versus unsuccessful Kirschstein-NRSA fellowship applicants) or groups that may reveal some interest (e.g., Kirschstein-NRSA trainees and other non-Kirschstein-NRSA postdoctorates).
Another type of comparison is between those with research doctorates and those with medical degrees, or those with dual (M.D. and Ph.D.) degrees. One question is whether those with medical degrees have the same career profiles as those with research doctorates. This question relates to the differing premises of the two types of degrees; that is, that medical training is oriented toward providing services to individuals, not toward conducting research.
Table 5.1 provides information on the six study groups, the outcomes that they inform, the data sources used, and the population.
Table 5.1. Summarization of Study Groups*
Study Group Degree Type
Relevant Outcomes Data Sources Population
Kirschstein-NRSA postdoctorates (fellows, trainees and trainee/fellows)
Ph.D. Grants CGAF Census of all individualsPublications ISI Subsample of SDR sampleEmployment SDR SDR sample
M.D./Ph.D. Grants CGAF Census of all individualsM.D. Grants CGAF Census of all individuals
Unsuccessful Kirschstein-NRSA fellowship applicants
Ph.D. Grants CGAF Census of all individualsPublications ISI Subsample of SDR sampleEmployment SDR SDR sample
Other non-Kirschstein-NRSA postdoctorates and other research doctorates
Ph.D. Grants CGAF SDR samplePublications ISI Subsample of SDR sampleEmployment SDR SDR sample
* The SDR is a sample of all research doctorates. From that sample, individuals were subsampled for the publications outcomes analysis.
The ability to report on all groups was largely dictated by the available data sources. The CGAF was available for all individuals, while the publications data supplied by ISI was available for a subsample of the SDR sample. The absolute number involved in the analysis of grants was large
ORC Macro 149
(about 39,000 individuals), while for the publications data it was relatively small (about 2,700 individuals).
5.2.1. Measuring the Effects of Outcomes
One critical issue in measuring difference in outcomes using the various study groups is selection. That is, to what extent do individual traits or predoctoral training experiences affect outcomes independent of Kirschstein-NRSA postdoctoral training? This issue would suggest that if Kirschstein-NRSA selected the best and brightest, any difference in outcomes may not be fully a result of postdoctoral training but may instead reflect selection factors. One approach to addressing this selection issue is to control for various critical factors. Chapter 2 describes factors that have been examined in the literature relative to whether they differentiate among the various study groups.
Chapter 2 provides an analysis of the following factors:
Research intensity of the institution granting the doctorate Elapsed and enrolled time in pursuit of the doctorate Receipt of Kirschstein-NRSA predoctoral training support Doctoral field Gender, marital status, and number of dependents Postdoctoral plans
In general, we found that there were large differences between Kirschstein-NRSA postdoctorates and other research doctorates (including non-Kirschstein-NRSA postdoctorates) on almost all of these measures. The factor that was most useful in distinguishing among the different Kirschstein-NRSA study groups was elapsed and enrolled time in pursuit of the doctorate. Analysis of that factor indicated that Kirschstein-NRSA postdoctoral fellows completed their doctoral requirements more quickly than trainees as well as all other study groups. This may indicate that such individuals may be unusual in their ability to advance their careers more quickly than those in the other study groups.
In chapters 2 and 3, we also compared Kirschstein-NRSA fellows and trainees by months of support. This comparison provided some indication that beyond the simple difference between Kirschstein-NRSA fellows and trainees, certain groups of fellows with more months of support were different from fellows with fewer months of support. This analysis provides some justification for examining the Kirschstein-NRSA postdoctoral population by months of support.
5.3. RESULTS OF THE OUTCOMES ANALYSIS
Chapter 4 presented the results of an analysis of outcomes focusing on the differences among the six study groups. The following results were observed relative to the three general outcome categories.
ORC Macro 150
5.3.1. Success of Research Doctorates in Obtaining NIH Research Grant Support
The first analysis focused only on research doctorates, primarily because it allows for comparison between Kirschstein-NRSA and non-Kirschstein-NRSA postdoctorate populations. The analysis focused on success in obtaining R01 and R01-equivalent funding. However, we also examined success in obtaining funding from more inclusive sets of grants that included R01 grants. In general, we found that the Kirschstein-NRSA postdoctoral population application rates and success rates were higher when these more inclusive categories were considered. Our conclusion was that Kirschstein-NRSA postdoctorates had more familiarity with or incentive to apply for these other grant mechanisms than other research doctorates.
Percentage of Postdoctorates Applying for R01 and R01-Equivalent Grants—The first measure in this category is defined as the percentage of individuals in the population who applied for these grants. The descriptive results for R01s show that Kirschstein-NRSA fellows and trainee/fellows had a 45-percent application rate, Kirschstein-NRSA trainees and unsuccessful Kirschstein-NRSA fellowship applicants had a 30-percent application rate, and the other two groups had a lower application rate. The regression analyses support these observed differences between Kirschstein-NRSA fellows and trainees and indicate a difference between Kirschstein-NRSA trainees and other non-Kirschstein-NRSA postdoctorates. When the R01 equivalent is considered, the application rates for Kirschstein-NRSA fellows and trainees increase more than for the non-Kirschstein-NRSA groups. For example, the rates for both fellows and trainees increase by 8 percent while the rate for other non-Kirschstein-NRSA postdoctorates increases by only 3 percent.
Percentage of Postdoctorates Awarded R01 and R01-Equivalent Grants—The second measure relates to the number of individuals who received an R01 or R01-equivalent grant out of the number who applied. The analysis indicates that Kirschstein-NRSA fellows stood out with a 62-percent success rate. Kirschstein-NRSA trainees, on the other hand, had a rate less than that of unsuccessful fellowship applicants and equal to that of other non-Kirschstein-NRSA postdoctorates. Again, success in being awarded the R01 equivalents favored Kirschstein-NRSA postdoctorates over the other groups, but not enough to change the results for trainees.
Number of R01 and R01-Equivalent Grants Awarded—Fellows have the highest number of grants, while trainees have an award total equivalent to other non-Kirschstein-NRSA postdoctorates. Unsuccessful Kirschstein-NRSA fellowship applicants were awarded more grants than Kirschstein-NRSA trainees. There is also some evidence that other Kirschstein-NRSA postdoctorates may perform as well as any other group within the first 10 years after they receive their degrees.
Award Rate for R01 and R01-Equivalent Grants—Award rates reflect the average number of applications awarded per individual. The data follow the trends that have been previously established, with Kirschstein-NRSA fellows exhibiting the highest rates and Kirschstein-NRSA trainee/fellows and unsuccessful Kirschstein-NRSA fellowship applicants showing the second-highest award rates. Other non-Kirschstein-NRSA postdoctorates have higher award rates than Kirschstein-NRSA trainees, with other postdoctorates having the lowest award rates. It should be noted that when R23 and R29 grants are also considered, trainees have an award rate
ORC Macro 151
equivalent to other postdoctorates, indicating that the R23 and R29 award mechanism may have favored trainees.
Years from Ph.D. to First Award—The last measure for this section reflects the time it takes to receive the first R01 and R01-equivalent award after award of the research doctorate. The data indicate that those who enter the job market earlier, other research doctorates, achieve their first R01 funding earlier. Kirschstein-NRSA fellows, other non-Kirschstein-NRSA postdoctorates, and unsuccessful fellowship applicants had similar results on this measure, and Kirschstein-NRSA trainees took the longest time to obtain funding. This finding was moderated when the R23 and R29 award mechanisms were considered: Trainees then had elapsed times equivalent to those of most groups, but not to fellows.
5.3.1.1. Overview
Two results emerge from this analysis. First, although there is a difference between the Kirschstein-NRSA-supported study groups and the non-Kirschstein-NRSA study groups, the difference reflects the performance of Kirschstein-NRSA fellows and trainee/fellows. In fact, trainees tend not to do as well as other Kirschstein-NRSA groups in securing R01 grant funding. Second, there is some evidence that the Kirschstein-NRSA-supported groups secure R01 and R01-equivalent funding later than other groups.
5.3.2. Success of All Kirschstein-NRSA Postdoctorates in Obtaining NIH Research Grant Support
This section discusses results relating to outcomes for all Kirschstein-NRSA postdoctorates (including professional doctorates) relating to securing research grant support, as follows:
Percentage Applying—The most notable result is that a higher percentage of Kirschstein-NRSA trainees and trainee/fellows with M.D.’s and M.D./Ph.D.’s apply for R01 grants. This result does not seem to apply to Kirschstein-NRSA fellows.
Percentage Receiving Awards—For Kirschstein-NRSA fellows, proportionally more Ph.D.’s and M.D./Ph.D.’s receive R01 funding than M.D.’s. However, for Kirschstein-NRSA trainees, M.D./Ph.D.’s are more successful than the other two degree groups in securing funding.
Number of Grants—Again, for Kirschstein-NRSA fellows, Ph.D.’s and M.D./Ph.D.’s receive more grants than those with M.D.’s. For Kirschstein-NRSA trainees, the M.D./Ph.D.’s outperformed those with other degrees.
Award Rate—For Kirschstein-NRSA fellows, Ph.D.’s and M.D./Ph.D.’s have higher R01 award rates. There is essentially no difference by degree group.
Average Years to First Award—For Kirschstein-NRSA fellows, Ph.D.’s and M.D./Ph.D.’s receive awards earlier than M.D.’s. For Kirschstein-NRSA trainees, M.D./Ph.D.’s take longer take longer than the other two degree groups.
ORC Macro 152
In summary, M.D.’s seem to have less success in securing R01 funding than Ph.D.’s and M.D./Ph.D.’s, particularly when the analysis focuses on Kirschstein-NRSA fellows. For Kirschstein-NRSA trainees, the M.D./Ph.D. group stands out in terms of securing R01 funding. It should be noted that there is a persistent difference across Kirschstein-NRSA support groups on these measures, suggesting that although degree makes a difference, the type of Kirschstein-NRSA support is critical to examining performance in terms of securing R01 funding.
5.3.3. Publication Success of Research Postdoctorates
Four measures of publication success were examined for two cohorts of research doctorates: 1981–82 and 1987–88 research doctorates. Publications were examined from the year individuals received their research doctorates until 1995. The analysis focused primarily on the 1981–82 cohort since they had a longer period in which to publish.
Percentage of the Study Groups Who Have Published—There was a difference between Kirschstein-NRSA postdoctoral groups and other study groups at various publication levels. Generally, there was little difference among the Kirschstein-NRSA postdoctoral study groups. There is some evidence that at a high threshold, 20 publications, Kirschstein-NRSA fellows were apt to be more prevalent than Kirschstein-NRSA trainees in the 1981–82 cohort, whereas in the 1987–88 cohort, trainees were slightly more prevalent. Individuals who were supported from 25 to 36 months were more likely to have published at the higher thresholds, whereas those who were supported for 0–9 months were least likely.
Number of Publications—Kirschstein-NRSA postdoctoral groups had more publications than other study groups. Regression results indicate that there was no difference among the Kirschstein-NRSA study groups. Again, those with 25–36 months of support had more publications than groups with other lengths of support.
Years to First Publication—Kirschstein-NRSA postdoctoral groups had the shortest elapsed time from the year they received their research doctorates until their first publication. There were no large differences among the Kirschstein-NRSA study groups. Again, those with 25–36 months of Kirschstein-NRSA support tended to publish earlier than other Kirschstein-NRSA postdoctorates.
Average Number of Citations Per Publication-Consistent with the other results, Kirschstein-NRSA postdoctoral study groups had a higher level of citations than the other study groups. Fellows showed a higher citation rate than trainees, and those with 25–36 months of support also showed a higher rate.
5.3.3.1. Overview
The analysis of publications shows a distinct difference between Kirschstein-NRSA postdoctorates and other research doctorates. Kirschstein-NRSA postdoctorates published more and sooner, and their publications were cited by other publications more often. There is very little evidence of differences among the Kirschstein-NRSA postdoctorate subgroups; however,
ORC Macro 153
there seems to be some evidence that fellows experience a lag in publishing when compared with trainees. Different lengths of support seem to be important: Kirschstein-NRSA postdoctorates with 25–36 months of support seem to have better publication records, while those with 0–9 months of support seem to have the worst publication records. It must be noted, however, that publications posed some challenges as an outcome measure. Publications may reflect group efforts rather than personal efforts and therefore may be difficult to use in examining individuals. Thus, an individual may get credit for a publication when in fact he or she made a relatively small contribution to the paper. Challenges in these counting procedures may account for some of the patterns seen in the results. In order to be sure, a greater understanding of the differences in the activities of fellows and trainees during their postdoctoral tenure should be explored.
5.3.4. Employment Success of Research Postdoctorates
The third set of outcomes relates to employment success, in particular outcomes that reflect a continuing interest and effort in research. The data to support this analysis were derived from the SDR. We assessed various employment-related measures collected through that longitudinal survey 4, 8, and 12 years after individuals received their doctorates.
Employment in Academic or Business/Industry Settings—Kirschstein-NRSA postdoctorates were more likely to find employment in academic settings than the other study groups. They were also more likely to be employed in academic settings than business/industry settings. After 12 years, however, they were more likely to be located in business/industry settings than the other study groups. Other research doctorates were more likely to be located outside these settings in government, the nonprofit sector, and other traditionally non-research-oriented settings.
Academic Rank and Tenure—Kirschstein-NRSA postdoctorates tend to assume associate and full professor positions more slowly than the other study groups. Even after 12 years, we see a small difference between groups, although the gap in the percentages has grown smaller. At the end of 12 years, a larger percentage of Kirschstein-NRSA fellows than Kirschstein-NRSA trainees attained associate and full professor positions. Results relating to the achievement of tenure status reflect these results.
5.3.4.1. Overview
The most critical result relating to the analysis of employment outcomes is that although Kirschstein-NRSA postdoctorates tend to focus on academic employment, their career in this setting is delayed. Other research doctorates, in particular, who obtain academic positions achieve tenure and higher academic rank in a shorter time after receiving their doctorates than Kirschstein-NRSA postdoctorates. Assuming that there are no differences in the inherent abilities of individuals who assume academic positions in these two groups, Kirschstein-NRSA postdoctoral training would initially seem to delay career development. However, it should be noted that Kirschstein-NRSA postdoctorates (and particularly fellows) seem to catch up with the other study groups. As implied, there is a difference between fellows and trainees, with fellows achieving higher academic rank more quickly. Finally, Kirschstein-NRSA postdoctoral fellows
ORC Macro 154
and trainees are more likely to be located in business and industry after 12 years, while other non-Kirschstein-NRSA postdoctorates tend to find themselves in other settings.
5.4. CONCLUSIONS
The major conclusion drawn from the study is that those in the Kirschstein-NRSA fellowship program seem to differentiate themselves from those in the other study groups, including Kirschstein-NRSA trainees. Such success, however, is delayed relative to those who begin their research careers directly after receiving their doctorates. The difference between Kirschstein-NRSA fellows and trainees is important to note since the outcomes of the latter group do not substantially differentiate themselves from other postdoctoral groups, such as those who applied unsuccessfully for a Kirschstein-NRSA fellowship. As chapter 3 discussed, the National Research Service Award program cannot be viewed as a unitary program because individuals take different postdoctoral support paths with differing lengths of support. Capturing these paths would provide a better sense of the kinds of research experiences provided and how they affect outcomes.
ORC Macro 155
APPENDIX A
PREDOCTORAL EXPERIENCES AND BACKGROUND CHARACTERISTICS OF NEW RESEARCH DOCTORATES TABLES
Table A.1. Number and Percentage of Research Doctorate Recipients, by Carnegie Classifications of the Universities Where They Received Their Doctorates and Type and Length of Postdoctoral Support
Type of Postdoctoral
Support
Length of Kirschstein-
NRSA Postdoctoral
Support (Months)
Carnegie Classification of InstitutionOther Universities TotalResearch
University 1Research
University 2Doctoral
University 1Doctoral
University 2
# % # % # % # % # % #All Kirschstein-NRSA postdoctorates
1–9 815 62.74 260 20.02 85 6.54 19 1.46 120 9.24 1,29910–12 2,826 64.76 764 17.51 339 7.77 79 1.81 356 8.16 4,36413–24 5,606 61.73 1,860 20.48 719 7.92 167 1.84 729 8.03 9,08125–36 4,727 61.18 1,576 20.40 603 7.80 113 1.46 707 9.15 7,72637+ 889 59.31 351 23.42 100 6.67 16 1.07 143 9.54 1,499Total 14,863 62.01 4,811 20.07 1,846 7.70 394 1.64 2,055 8.57 23,969
Kirschstein-NRSA fellows
1–9 226 66.86 63 18.64 17 5.03 4 1.18 28 8.28 33810–12 610 66.23 159 17.26 76 8.25 17 1.85 59 6.41 92113–24 2,277 63.93 712 19.99 251 7.05 44 1.24 278 7.80 3,56225–36 1,779 67.39 495 18.75 162 6.14 21 0.80 183 6.93 2,64037+ 175 67.57 48 18.53 20 7.72 2 0.77 14 5.41 259Total 5,067 65.63 1,477 19.13 526 6.81 88 1.14 562 7.28 7,720
Kirschstein-NRSA trainees
1–9 586 61.17 197 20.56 68 7.10 15 1.57 92 9.60 95810–12 2,212 64.42 603 17.56 261 7.60 62 1.81 296 8.62 3,43413–24 3,150 60.08 1,094 20.87 447 8.53 120 2.29 432 8.24 5,24325–36 2,151 57.59 796 21.31 329 8.81 81 2.17 378 10.12 3,73537+ 285 54.39 150 28.63 31 5.92 8 1.53 50 9.54 524Total 8,384 60.34 2,840 20.44 1,136 8.18 286 2.06 1,248 8.98 13,894
Kirschstein-NRSA trainee/fellows
1–9 3 100.00 0 0.00 0 0.00 0 0 0 0.00 310–12 4 44.44 2 22.22 2 22.22 0 0 1 11.11 913–24 179 64.86 54 19.57 21 7.61 3 1.09 19 6.88 27625–36 797 58.99 285 21.10 112 8.29 11 0.81 146 10.81 1,35137+ 429 59.92 153 21.37 49 6.84 6 0.84 79 11.03 716Total 1,412 59.96 494 20.98 184 7.81 20 0.85 245 10.40 2,355
Unsuccessful fellowship applicants 6,159 63.33 1,915 19.69 790 8.12 169 1.74 692 7.12 9,725Other postdoctorates 4,237 51.90 1,785 21.86 1,063 13.02 397 4.86 682 8.35 8,164Source: DRF
Table A.2. Number and Percentage of Research Doctorate Recipients, by Ranking of the Institution Where They Received Their Doctorates and Type and Length of Postdoctoral Support
Type of Postdoctoral Support
Months of Support
Ranking of Institutions Conferring the Doctorate*Other Institutions Total
Institutions Ranked in the Top 25
Institutions Ranked 26–100
# % # % # % #All Kirschstein-NRSA postdoctorates
1–9 454 34.9 397 30.6 448 34.5 1,29910–12 1,532 35.1 1,402 32.1 1,430 32.8 4,36413–24 3,219 35.4 2,836 31.2 3,026 33.3 9,08125–36 2,727 35.3 2,468 31.9 2,531 32.8 7,72637+ 520 34.7 485 32.4 494 33.0 1,499Total 8,452 35.3 7,588 31.7 7,929 33.1 23,969
Kirschstein-NRSA fellows 1–9 119 35.2 107 31.7 112 33.1 33810–12 325 35.3 293 31.8 303 32.9 92113–24 1,311 36.8 1,102 30.9 1,149 32.3 3,56225–36 1,017 38.5 844 32 779 29.5 2,64037+ 88 34.0 88 34.0 83 32 259Total 2,860 37.0 2,434 31.5 2426 31.4 7,720
Kirschstein-NRSA trainees
1–9 333 34.8 289 30.2 336 35.1 95810–12 1,204 35.1 1,106 32.2 1,124 32.7 3,43413–24 1,806 34.4 1,636 31.2 1,801 34.4 5,24325–36 1,237 33.1 1,198 32.1 1,300 34.8 3,73537+ 192 36.6 163 31.1 169 32.3 524Total 4,772 34.3 4,392 31.6 4,730 34.0 13,894
Kirschstein-NRSA trainee/fellows
1–9 2 66.7 1 33.3 0 0.0 310–12 3 33.3 3 33.3 3 33.3 913–24 102 37.0 98 35.5 76 27.5 27625–36 473 35.0 426 31.5 452 33.5 1,35137+ 240 33.5 234 32.7 242 33.8 716Total 820 34.8 762 32.4 773 32.8 2,355
Unsuccessful fellowship applicants 3,503 36 2,958 30.4 3,264 33.6 9,725Other research doctorates 2,541 31.1 2,180 26.7 3,443 42.2 8,164* The rankings are as of the date of receipt of the doctorate.Sources: CGAF and DRF
Table A.3. Number and Percentage of Research Doctorates, by Type of NIH Predoctoral Support and Type and Length of Postdoctoral Support
Type of Postdoctoral Support
Months of Kirschstein-NRSA
Postdoctoral Support
Predoctoral SupportTotalKirschstein-NRSA
Support Other NIH Support No Support Indicated
# % # % # % #All Kirschstein-NRSA postdoctorates
1-9 338 26.0 203 15.6 758 58.4 1,29910–12 1,020 23.4 656 15.0 2,688 61.6 4,36413–24 1,969 21.7 1,478 16.3 5,634 62.0 9,08125–36 2,150 27.8 1,200 15.5 4,376 56.6 7,72637+ 365 24.3 229 15.3 905 60.4 1,499Total 5,842 24.4 3,766 15.7 14,361 59.9 23,969
Kirschstein-NRSA fellows
1–9 81 24.0 61 18.0 196 58.0 33810–12 162 17.6 185 20.1 574 62.3 92113–24 797 22.4 700 19.7 2,065 58.0 3,56225–36 877 33.2 458 17.3 1,305 49.4 2,64037+ 76 29.3 52 20.1 131 50.6 259Total 1,993 25.8 1,456 18.9 4,271 55.3 7,720
Kirschstein-NRSA trainees
1–9 256 26.7 142 14.8 560 58.5 95810–12 854 24.9 471 13.7 2,109 61.4 3,43413–24 1,089 20.8 741 14.1 3,413 65.1 5,24325–36 869 23.3 518 13.9 2,348 62.9 3,73537+ 98 18.7 66 12.6 360 68.7 524Total 3,166 22.8 1,938 13.9 8,790 63.3 13,894
Kirschstein-NRSA trainee/fellows
1–9 1 33.3 0 0.0 2 66.7 310–12 4 44.4 0 0.0 5 55.6 913–24 83 30.1 37 13.4 156 56.5 27625–36 404 29.9 224 16.6 723 53.5 1,35137+ 191 26.7 111 15.5 414 57.8 716Total 683 29.0 372 15.8 1,300 55.2 2,355
Unsuccessful fellowship applicants 2,153 22.2 1,634 16.9 5,904 60.9 9,691
Sources: DRF, TFF
Table A.4. Distribution and Median Number of Months Elapsed From First Enrollment in Graduate School to Receipt of the Doctorate, by Type and Length of Postdoctoral Support
Type of Postdoctoral
Support
Months of Kirschstein-
NRSA Postdoctoral
Support
Distribution and Median Number of Months Spent in Pursuit of the DoctorateTotal48 or Fewer
Months 49–60 Months 61–72 Months 73–84 Months 85+ Months Unknown Months*
# % # % # % # % # % # % #All Kirschstein-NRSA postdoctorates
1–9 80 6.2 219 16.9 255 19.6 180 13.9 468 36.0 97 7.5 1,29910–12 256 5.9 640 14.7 833 19.1 626 14.3 1,664 38.1 345 7.9 4,36413–24 561 6.2 1,450 16.0 1,849 20.4 1,298 14.3 3,371 37.1 552 6.1 9,08125–36 337 4.4 1,119 14.5 1,574 20.4 1,233 16.0 2,987 38.7 476 6.2 7,72637+ 52 3.5 201 13.4 308 20.5 231 15.4 594 39.6 113 7.5 1,499Total 1,286 5.4 3,629 15.1 4,819 20.1 3,568 14.9 9,084 37.9 1,583 6.6 23,969
Kirschstein-NRSA fellows
1–9 31 9.2 83 24.6 85 25.1 39 11.5 83 24.6 17 5.0 33810–12 86 9.3 187 20.3 198 21.5 119 12.9 266 28.9 65 7.1 92113–24 266 7.5 747 21.0 827 23.2 512 14.4 1,029 28.9 181 5.1 3,56225–36 122 4.6 450 17.0 618 23.4 448 17.0 872 33.0 130 4.9 2,64037+ 9 3.5 39 15.1 62 23.9 47 18.1 84 32.4 18 6.9 259Total 514 6.7 1,506 19.5 1,790 23.2 1,165 15.1 2,334 30.2 411 5.3 7,720
Kirschstein-NRSA trainees
1–9 49 5.1 135 14.1 169 17.6 141 14.7 384 40.1 80 8.4 95810–12 169 4.9 452 13.2 633 18.4 505 14.7 1,395 40.6 280 8.2 3,43413–24 267 5.1 651 12.4 962 18.3 742 14.2 2,267 43.2 354 6.8 5,24325–36 127 3.4 444 11.9 658 17.6 579 15.5 1,663 44.5 264 7.1 3,73537+ 15 2.9 44 8.4 82 15.6 85 16.2 247 47.1 51 9.7 524Total 627 4.5 1,726 12.4 2,504 18.0 2,052 14.8 5,956 42.9 1,029 7.4 13,894
Kirschstein-NRSA trainee/fellows
1–9 0 0.0 1 33.3 1 33.3 0 0.0 1 33.3 0 0.0 310–12 1 11.1 1 11.1 2 22.2 2 22.2 3 33.3 0 0.0 913–24 28 10.1 52 18.8 60 21.7 44 15.9 75 27.2 17 6.2 27625–36 88 6.5 225 16.7 298 22.1 206 15.2 452 33.5 82 6.1 1,35137+ 28 3.9 118 16.5 164 22.9 99 13.8 263 36.7 44 6.1 716Total 145 6.2 397 16.9 525 22.3 351 14.9 794 33.7 143 6.1 2,355
Unsuccessful fellowship applicants 546 5.6 1,647 16.9 2,009 20.7 1,461 15.0 3,442 35.4 620 6.4 9,725
Other research doctorates 354 4.3 827 10.1 1,003 12.3 839 10.3 4,241 51.9 900 11.0 8,164* These were omitted from the percentages presented in chapter 2.Source: DRF
Table A.5. Distribution and Median Number of Months Enrolled in Graduate School to Receipt of the Doctorate, by Type and Length of Postdoctoral Support
Type of Postdoctoral Support
Months of Kirschstein-
NRSA Postdoctoral
Support
Distribution and Median Number of Months Spent in Pursuit of the DoctorateTotal48 or Fewer
Months 49–60 Months 61–72 Months 73–84 Months 85+ Months Unknown Months*
# % # % # % # % # % # % #All Kirschstein-NRSA postdoctorates
1–9 182 14.0 310 23.9 274 21.1 184 14.2 191 14.7 158 12.2 1,29910–12 506 11.6 930 21.3 953 21.8 616 14.1 781 17.9 578 13.2 4,36413–24 1,172 12.9 2,029 22.3 2,119 23.3 1,251 13.8 1,524 16.8 986 10.9 9,08125–36 803 10.4 1,612 20.9 1,866 24.2 1,145 14.8 1,457 18.9 843 10.9 7,72637+ 143 9.5 297 19.8 355 23.7 207 13.8 314 20.9 183 12.2 1,499Total 2,806 11.7 5,178 21.6 5,567 23.2 3,403 14.2 4,267 17.8 2,748 11.5 23,969
Kirschstein-NRSA fellows
1–9 65 19.2 108 32.0 68 20.1 29 8.6 37 10.9 31 9.2 33810–12 167 18.1 236 25.6 193 21.0 108 11.7 116 12.6 101 11.0 92113–24 516 14.5 999 28.0 860 24.1 448 12.6 401 11.3 338 9.5 3,56225–36 308 11.7 605 22.9 722 27.3 364 13.8 390 14.8 251 9.5 2,64037+ 33 12.7 58 22.4 62 23.9 38 14.7 42 16.2 26 10.0 259Total 1,089 14.1 2,006 26.0 1,905 24.7 987 12.8 986 12.8 747 9.7 7,720
Kirschstein-NRSA trainees
1–9 115 12.0 202 21.1 205 21.4 155 16.2 154 16.1 127 13.3 95810–12 337 9.8 693 20.2 758 22.1 506 14.7 664 19.3 476 13.9 3,43413–24 606 11.6 960 18.3 1,200 22.9 765 14.6 1,094 20.9 618 11.8 5,24325–36 313 8.4 718 19.2 817 21.9 579 15.5 859 23.0 449 12.0 3,73537+ 38 7.3 81 15.5 107 20.4 79 15.1 139 26.5 80 15.3 524Total 1,409 10.1 2,654 19.1 3,087 22.2 2,084 15.0 2,910 20.9 1,750 12.6 13,894
Kirschstein-NRSA trainee/fellows
1–9 2 66.7 0 0.0 1 33.3 0 0.0 0 0.0 0 0.0 310–12 2 22.2 1 11.1 2 22.2 2 22.2 1 11.1 1 11.1 913–24 50 18.1 70 25.4 59 21.4 38 13.8 29 10.5 30 10.9 27625–36 182 13.5 289 21.4 327 24.2 202 15.0 208 15.4 143 10.6 1,35137+ 72 10.1 158 22.1 186 26.0 90 12.6 133 18.6 77 10.8 716Total 308 13.1 518 22.0 575 24.4 332 14.1 371 15.8 251 10.7 2,355
Unsuccessful fellowship applicants 1,189 12.2 2,286 23.5 2,134 21.9 1,343 13.8 1,622 16.7 1,151 11.8 9,725Other research doctorates 894 11.0 1,309 16.0 1,395 17.1 1,074 13.2 2,197 26.9 1,295 15.9 8,164* These were omitted from the percentages presented in chapter 2.Source: DRFTable A.6. Number and Percentage of Research Doctorates in Kirschstein-NRSA Programs, by Gender
Male Female Total
Type of Postdoctoral
Support
Months of Kirschstein-NRSA
Postdoctoral Support
# % # % #
Kirschstein-NRSA fellows
1–9 256 75.7 82 24.3 33810–12 668 72.6 252 27.4 92013–24 2,536 71.2 1,024 28.8 3,56025–36 1,710 64.8 929 35.2 2,63937+ 139 53.7 120 46.3 259Total 5,309 68.8 2,407 31.2 7,716
Kirschstein-NRSA trainees
1–9 603 62.9 355 37.1 95810–12 2,212 64.4 1,222 35.6 3,43413–24 3,282 62.6 1,959 37.4 5,24125–36 2,477 66.4 1,254 33.6 3,73137+ 364 69.5 160 30.5 524Total 8,938 64.4 4,950 35.6 13,888
Kirschstein-NRSA trainee/fellows
1–9 3 100.0 0 0.0 310–12 5 55.6 4 44.4 913–24 190 68.8 86 31.2 27625–36 888 65.8 462 34.2 1,35037+ 463 64.7 253 35.3 716Total 1,549 65.8 805 34.2 2,354
Sources: DRF, TFF
Table A.7. Number and Percentage of Research Doctorates, by Marital Status at Receipt of Doctorate and Type and Length of Kirschstein-NRSA Postdoctoral Support
Type of Postdoctoral
Support
Months of Kirschstein-
NRSA Postdoctoral
Support
Married Unmarried Unknown* Total
# % # % # % #Kirschstein-NRSA fellows
1–9 186 55.0 135 39.9 17 5.0 33810–12 510 55.4 382 41.5 29 3.1 92113–24 1,930 54.2 1,497 42.0 135 3.8 3,56225–36 1,263 47.8 1,267 48.0 110 4.2 2,64037+ 126 48.6 121 46.7 12 4.6 259Total 4,015 52.0 3,402 44.1 303 3.9 7,720
Kirschstein-NRSA trainees
1–9 524 54.7 394 41.1 40 4.2 95810–12 1,757 51.2 1,500 43.7 177 5.2 3,43413–24 2,684 51.2 2,330 44.4 229 4.4 5,24325–36 1,905 51.0 1,664 44.6 166 4.4 3,73537+ 263 50.2 238 45.4 23 4.4 524Total 7,133 51.3 6,126 44.1 635 4.6 13,894
Kirschstein-NRSA trainee/fellows
1–9 1 33.3 2 66.7 0 0.0 310–12 8 88.9 1 11.1 0 0.0 913–24 137 49.6 125 45.3 14 5.1 27625–36 667 49.4 630 46.6 54 4.0 1,35137+ 337 47.1 358 50.0 21 2.9 716Total 1,150 48.8 1,116 47.4 89 3.8 2,355
* These were omitted from the percentages presented in chapter 2.Sources: DRF, TFF
Table A.8. Number and Percentage of Research Doctorates, by Number of Dependents and Type and Length of Postdoctoral Support
Type of Postdoctoral
Support
Months of Kirschstein-
NRSA Postdoctoral
Support
Number of DependentsTotal
No Dependents One Dependent Two of More Dependents
Unknown Dependents*
# % # % # % # % #All Kirschstein-NRSA postdoctorates
1–9 648 49.9 238 18.3 197 15.2 216 16.6 1,29910–12 2,285 52.4 700 16.0 653 15.0 726 16.6 4,36413–24 4,903 54.0 1,490 16.4 1,221 13.4 1,467 16.2 9,08125–36 4,468 57.8 1,217 15.8 913 11.8 1,128 14.6 7,72637+ 878 58.6 248 16.5 189 12.6 184 12.3 1,499Total 13,182 55.0 3,893 16.2 3,173 13.2 3,721 15.5 23,969
Kirschstein-NRSA fellows
1–9 178 52.7 52 15.4 52 15.4 56 16.6 33810–12 453 49.2 169 18.3 137 14.9 162 17.6 92113–24 1,891 53.1 621 17.4 438 12.3 612 17.2 3,56225–36 1,633 61.9 380 14.4 247 9.4 380 14.4 2,64037+ 162 62.6 33 12.7 30 11.6 34 13.1 259Total 4,317 55.9 1,255 16.3 904 11.7 1,244 16.1 7,720
Kirschstein-NRSA trainees
1–9 468 48.8 186 19.4 144 15.0 160 16.7 95810–12 1,829 53.3 528 15.4 513 14.9 564 16.4 3,43413–24 2,864 54.6 825 15.7 758 14.5 796 15.2 5,24325–36 2,030 54.3 635 17.0 519 13.9 551 14.7 3,73537+ 292 55.7 94 17.9 77 14.7 61 11.6 524Total 7,483 53.9 2,268 16.3 2,011 14.5 2,132 15.3 13,894
Kirschstein-NRSA trainee/fellows
1–9 2 66.7 0 0.0 1 33.3 0 0.0 310–12 3 33.3 3 33.3 3 33.3 0 0.0 913–24 148 53.6 44 15.9 25 9.1 59 21.4 27625–36 805 59.6 202 14.9 147 10.9 197 14.6 1,35137+ 424 59.2 121 16.9 82 11.4 89 12.4 716Total 1,382 58.7 370 15.7 258 11.0 345 14.6 2,355
Unsuccessful fellowship applicants 5,102 52.5 1,578 16.2 1,331 13.7 1,714 17.6 9,725Other research doctorates 3,293 40.3 1,328 16.3 2,143 26.2 1,400 17.1 8,164* These were omitted from the percentages presented in chapter 2.Source: DRF
Table A.9. Number and Percentage of New Doctorate Recipients, by Postdoctoral Plans and Type and Length of Postdoctoral Support
Type of Postdoctoral Support
Months of Kirschstein-
NRSA Postdoctoral
Support
Postdoctoral PlansTotal
Had Postdoctoral Plans Did not Have Postdoctoral Plans
Unknown Postdoctoral Plans
# % # % # % #All Kirschstein-NRSA postdoctorates
1–9 1,025 78.9 214 16.5 60 4.6 1,29910–12 3,339 76.5 792 18.1 233 5.3 4,36413–24 7,570 83.7 1,082 11.9 429 4.7 9,08125–36 6,766 87.6 621 8.0 339 4.4 7,72637+ 1,309 87.3 125 8.3 65 4.3 1,499Total 20,009 83.5 2,834 11.8 1,126 4.7 23,969
Kirschstein-NRSA fellows 1–9 298 88.2 25 7.4 15 4.4 33810–12 771 83.7 117 12.7 33 3.6 92113–24 3,201 89.9 216 6.1 145 4.1 3,56225–36 2,423 91.8 114 4.3 103 3.9 2,64037+ 226 87.3 21 8.1 12 4.6 259Total 6,919 89.6 493 6.4 308 4.0 7,720
Kirschstein-NRSA trainees
1–9 725 75.7 188 19.6 45 4.7 95810–12 2,560 74.5 674 19.6 200 5.8 3,43413–24 4,135 78.9 841 16.0 267 5.1 5,24325–36 3,100 83.0 454 12.2 181 4.8 3,73537+ 429 81.9 66 12.6 29 5.5 524Total 10,949 78.8 2,223 16.0 722 5.2 13,894
Kirschstein-NRSA trainee/fellows
1–9 2 66.7 1 33.3 0 0 310–12 8 88.9 1 11.1 0 0 913–24 234 84.8 25 9.6 17 6.2 27625–36 1,243 92.0 53 3.9 55 4.1 1,35137+ 654 91.3 38 5.3 24 3.4 716Total 2,141 90.9 118 5.0 96 4.1 2,355
Unsuccessful fellowship applicants 7,651 78.7 1,604 16.5 470 4.8 9,725Other research doctorates 2,265 27.7 5,676 69.5 223 2.7 8,164
Source: DRF
APPENDIX B
NATURE OF KIRSCHSTEIN-NRSA SUPPORT TABLES
Table B.1. Length of Initial Kirschstein-NRSA Competing Fellowship, by Type of Degree
Total Months of Support Per Competing Fellowship Award
Ph.D. M.D. M.D./Ph.D. Total
# % # % # % # %<=12 months 2,324 25.2 427 37.3 180 23.8 2,931 26.413 to 24 months 3,856 41.8 520 45.4 308 40.7 4,684 42.125 to 36 months 2,405 26.1 160 14.0 221 29.2 2,786 25.0> 36 months 636 6.9 38 3.3 48 6.3 722 6.5Total 9,221 100.0 1,145 100.0 757 100.0 11,123 100.0Notes: Counts are for the first F32 and F35 new and competing grants for those fellowship applicants in the Postdoc file. The grants period is between 1975 and 1992.
Table B.2. Number and Percentage of Kirschstein-NRSA Postdoctorates, by Type of Support, Months of Postdoctoral Support, and Type of Degree
Type ofRecipient
Months ofSupport
Type of DegreeTotal
Ph.D. M.D. M.D./Ph.D.
# % # % # % # %
All Kirschstein-NRSA postdoctorates
1–9 1,198 5.6 217 3.1 101 4.2 1,516 4.910–12 3,803 17.6 2,484 36.0 561 23.4 6,848 22.213–24 8,291 38.4 2,683 38.9 790 33.0 11,764 38.125–36 7,009 32.5 1,295 18.8 717 29.9 9,021 29.237+ 1,271 5.9 212 3.1 228 9.5 1,711 5.5Total 21,572 100.0 6,891 100.0 2,397 100.0 30,860 100.0
Kirschstein-NRSA fellows 1–9 310 4.3 29 4.4 28 5.4 367 4.410–12 865 12.0 152 22.8 56 10.9 1,073 12.813–24 3,344 46.4 340 51.1 218 42.3 3,902 46.525–36 2,451 34.0 138 20.7 189 36.7 2,778 33.137+ 235 3.3 7 1.1 24 4.7 266 3.2Total 7,205 100.0 666 100.0 515 100.0 8,386 100.0
Kirschstein-NRSA trainees 1–9 885 7.2 188 3.3 73 4.5 1,146 5.810–12 2,929 23.9 2,328 40.6 505 30.9 5,762 29.413–24 4,703 38.4 2,226 38.8 540 33.1 7,469 38.125–36 3,334 27.2 890 15.5 401 24.6 4,625 23.637+ 411 3.4 100 1.7 113 6.9 624 3.2Total 12,262 100.0 5,732 100.0 1,632 100.0 19,626 100.0
Kirschstein-NRSA trainee/fellows
1–9 3 0.1 0 0.0 0 0.0 3 0.110–12 9 0.4 4 0.8 0 0.0 13 0.513–24 244 11.6 117 23.7 32 12.8 393 13.825–36 1,224 58.1 267 54.2 127 50.8 1,618 56.837+ 625 29.7 105 21.3 91 36.4 821 28.8Total 2,105 100.0 493 100.0 250 100.0 2,848 100.0
Table B.3. Number and Percentage of Kirschstein-NRSA Fellows and Trainees, by Area of Research, Type of Kirschstein-NRSA Postdoctoral Support, and Degree
Type of Postdoctoral
SupportDiscipline
DegreeTotal
Ph.D. M.D. M.D./Ph.D. Other# % # % # % # % #
All Kirschstein-NRSA postdoctorates
Biomedical sciences 11,424 81.5 1,007 7.2 865 6.2 729 5.2 14,025Chemistry 619 92.1 8 1.2 13 1.9 32 4.8 672Clinical sciences 1,387 19.3 4,848 67.4 861 12.0 99 1.4 7,195Behavioral and social sciences 1,855 90.0 42 2.0 41 2.0 123 6.0 2,061Other disciplines 2,359 75.2 280 8.9 220 7.0 278 8.9 3,137Unknown 3,928 73.2 706 13.2 397 7.4 336 6.3 5,367
Kirschstein-NRSA fellows
Biomedical sciences 4,030 88.3 181 4.0 214 4.7 137 3.0 4,562Chemistry 408 95.1 1 0.2 10 2.3 10 2.3 429Clinical sciences 198 35.1 266 47.2 89 15.8 11 2.0 564Behavioral and social sciences 417 91.6 4 0.9 11 2.4 23 5.1 455Other disciplines 155 78.3 6 3.0 25 12.6 12 6.1 198Unknown 1,997 79.3 208 8.3 166 6.6 147 5.8 2,518
Kirschstein-NRSA trainees
Biomedical sciences 6,161 77.6 690 8.7 550 6.9 543 6.8 7,944Chemistry 194 86.2 7 3.1 3 1.3 21 9.3 225Clinical sciences 1,128 17.8 4,398 69.4 721 11.4 86 1.4 6,333Behavioral and social sciences 1,374 89.3 38 2.5 29 1.9 97 6.3 1,538Other disciplines 2,087 75.0 258 9.3 178 6.4 260 9.3 2,783Unknown 1,318 66.8 341 17.3 151 7.6 164 8.3 1,974
Kirschstein-NRSA trainee/fellows
Biomedical sciences 1,233 81.2 136 9.0 101 6.6 49 3.2 1,519Chemistry 17 94.4 0 0.0 0 0.0 1 5.6 18Clinical sciences 61 20.5 184 61.7 51 17.1 2 0.7 298Behavioral and social sciences 64 94.1 0 0.0 1 1.5 3 4.4 68Other disciplines 117 75.0 16 10.3 17 10.9 6 3.8 156Unknown 613 70.1 157 17.9 80 9.1 25 2.9 875
Total 21,572 66.5 6,891 21.2 2,397 7.4 1,597 4.9 32,457
Table B.4. Number and Percentage of Kirschstein-NRSA Fellows and Trainees, by Carnegie Classification of Institutions and Type and Length of Postdoctoral Support
Type of Postdoctoral
SupportMonths of Support
Carnegie Classification of Postdoctoral InstitutionOther
Universities TotalResearch University 1
Research University 2
Doctoral University 1
Doctoral University 2
# % # % # % # % # % #All Kirschstein-NRSA postdoctorates
1–9 905 54.8 145 8.8 35 2.1 8 0.5 558 33.8 1,65110–12 4,110 56.6 392 5.4 127 1.7 38 0.5 2,595 35.7 7,26213–24 7,082 57.5 783 6.4 235 1.9 58 0.5 4,157 33.8 12,31525–36 5,342 56.6 500 5.3 155 1.6 53 0.6 3,381 35.8 9,43137+ 956 53.2 87 4.8 35 1.9 17 0.9 703 39.1 1,798Total 18,395 56.7 1,907 5.9 587 1.8 174 0.5 11,394 35.1 32,457
Kirschstein-NRSA fellows
1–9 213 54.1 22 5.6 10 2.5 2 0.5 147 37.3 39410–12 594 52.5 68 6.0 35 3.1 1 0.1 434 38.3 1,13213–24 2,242 55.4 281 6.9 91 2.2 15 0.4 1,419 35.1 4,04825–36 1,613 56.1 164 5.7 52 1.8 10 0.3 1,036 36.0 2,87537+ 124 44.8 16 5.8 3 1.1 4 1.4 130 46.9 277Total 4,786 54.8 551 6.3 191 2.2 32 0.4 3,166 36.3 8,726
Kirschstein-NRSA trainees
1–9 690 55.1 122 9.7 25 2.0 6 0.5 410 32.7 1,25310–12 3,511 57.4 323 5.3 92 1.5 37 0.6 2,154 35.2 6,11713–24 4,624 58.8 477 6.1 133 1.7 42 0.5 2,587 32.9 7,86325–36 2,760 56.4 248 5.1 88 1.8 33 0.7 1,768 36.1 4,89737+ 373 55.9 26 3.9 17 2.5 6 0.9 245 36.7 667Total 11,958 57.5 1,196 5.8 355 1.7 124 0.6 7,164 34.4 20,797
Kirschstein-NRSA trainee/fellows
1–9 2 50.0 1 25.0 0 0.0 0 0.0 1 25.0 410–12 5 38.5 1 7.7 0 0.0 0 0.0 7 53.8 1313–24 216 53.5 25 6.2 11 2.7 1 0.2 151 37.4 40425–36 969 58.4 88 5.3 15 0.9 10 0.6 577 34.8 1,65937+ 459 53.7 45 5.3 15 1.8 7 0.8 328 38.4 854Total 1,651 56.3 160 5.5 41 1.4 18 0.6 1,064 36.3 2,934
Table B.5. Number and Percentage of Kirschstein-NRSA Postdoctorates with Postdoctoral Support at Institutions Ranked by NIH Extramural Support Received, by Type and Length of Postdoctoral Support
Type of Postdoctoral Support
Months of Postdoctoral
Support
Postdoctoral Support Institutions*TotalInstitutions Ranked in the
Top 25Institutions Ranked
26–100 Other Institutions
# % # % # % #All Kirschstein-NRSA postdoctorates
1–9 761 46.1 614 37.2 276 16.7 1,65110–12 3,767 51.9 2,414 33.2 1,081 14.9 7,26213–24 6,155 50.0 4,122 33.5 2,038 16.5 12,31525–36 4,630 49.1 3,295 34.9 1,506 16.0 9,43137+ 873 48.6 636 35.4 289 16.1 1,798Total 16,186 49.9 11,081 34.1 5,190 16.0 32,457
Kirschstein-NRSA fellows
1–9 144 36.5 142 36.0 108 27.4 39410–12 424 37.5 407 36.0 301 26.6 1,13213–24 1,583 39.1 1,482 36.6 983 24.3 4,04825–36 1,119 38.9 1,040 36.2 716 24.9 2,87537+ 80 28.9 111 40.1 86 31.0 277Total 3,350 38.4 3,182 36.5 2,194 25.1 8,726
Kirschstein-NRSA trainees
1–9 616 49.2 469 37.4 168 13.4 1,25310–12 3,338 54.6 2,000 32.7 779 12.7 6,11713–24 4,367 55.5 2,502 31.8 994 12.6 7,86325–36 2,604 53.2 1,705 34.8 588 12.0 4,89737+ 348 52.2 234 35.1 85 12.7 667Total 11,273 54.2 6,910 33.2 2,614 33.7 20,797
Kirschstein-NRSA trainee/fellows
1–9 1 25.0 3 75.0 0 0.0 410–12 5 38.5 7 53.8 1 7.7 1313–24 205 50.7 138 34.2 61 15.1 40425-36 907 54.7 550 33.2 202 12.2 1,65937+ 445 52.1 291 34.1 118 13.8 854Total 1,563 53.3 989 33.7 382 13.0 2,934
* At the time of the postdoctoral supportSources: CGAF, TFF
APPENDIX C
CAREER OUTCOMES TABLES
GRANT OUTCOMES
These tables represent all Kirschstein-NRSA research postdoctorates (Ph.D.’s) who met criteria related to degree year and support year, as well as all unsuccessful fellowship applicants who met the same criteria. The number of observations represents a census derived from the TFF. In addition to these study groups, two other Ph.D. study groups are included. Individuals in these study groups include those who responded to the 1995 Survey of Doctorate Recipients and who met the various criteria related to award of degree and field of study. The first of these groups indicated that they were involved in postdoctoral study other than that supported by Kirschstein-NRSA, while the second group did not indicate any postdoctoral study.
Table C.1. Number and Percentage of Ph.D.’s Submitting a Competing NIH Grant Application Within the First 10 Years After Receiving the Doctorate and Percentage of Grant Applicants Awarded For Selected NIH Grant Mechanisms, by Subgroup
Study Group All Research Postdoctorates
Percentage of Research Doctorates Submitting an Application for an NIH Grant in the First 10 years After Receiving the Doctorate
R01R01, R23, and
R29 (Equivalents)
All RPGs Early K Awards*
Late K Awards**
K Awards Total KRUMPT
# % # % # % # % # % # % # %Kirschstein-NRSA fellows
7,205 3,241 45.0 3,808 52.9 3,942 54.7 25 0.4 469 6.5 497 6.9 4.075 56.6
Kirschstein-NRSA trainees
12,262 3,615 29.5 4,640 37.8 4,988 40.7 89 0.7 324 2.6 431 3.5 5,256 42.9
Kirschstein-NRSA trainee/fellows
2,105 973 46.2 1,205 57.2 1,250 59.4 14 0.7 116 5.5 131 6.2 1,292 61.4
Unsuccessful fellowship applicants
9,516 2,896 30.4 3,357 35.3 3,590 37.7 74 0.8 335 3.5 426 4.5 3.815 40.1
Other postdoctorates
2,432 480 19.7 547 22.5 606 24.9 17 0.7 60 2.5 81 3.3 652 26.8
Other research doctorates
5,535 517 9.3 557 10.1 674 12.2 23 0.4 36 0.7 68 1.2 771 13.9
Source: CGAF
Table C.2. Number and Percentage of Ph.D.’s Submitting a Competing NIH Grant Application Within the First 10 Years After Receiving the Doctorate Who Were Awarded a Grant for Selected NIH Grant Mechanisms, by Subgroup
Study Group
Percentage of those Submitting in the First 10 Years Who Receive Awards*R01 R01, R23, and
R29 (Equivalents)
All RPGs Early K Awards
Late K Awards K Awards KRUMPT
# % # % # % # % # % # % # %Kirschstein-NRSA fellows 1,990 61.3 2,544 66.7 2,706 68.7 9 36.0 222 47.3 233 46.4 2,807 68.9Kirschstein-NRSA trainees 1,686 46.8 2,437 52.6 2,715 54.3 48 53.9 129 39.8 181 42.9 2,902 55.2Kirschstein-NRSA trainee/fellows
553 56.9 780 64.9 824 65.8 10 71.4 51 44.0 62 48.4 864 66.9
Unsuccessful fellowship applicants
1,550 53.6 1,926 57.2 2,125 59.2 34 45.9 143 42.7 187 44.4 2,279 59.9
Other postdoctorates 237 49.4 279 51.0 323 53.3 9 52.9 27 45.0 38 46.9 356 54.6Other research doctorates 184 35.6 210 37.7 276 40.9 9 39.1 13 36.1 26 38.2 325 42.5* The denominators for the percentages are provided in table C.1 as cell observations.Source: CGAF
Table C.3. Average Award Rate* of Those Submitting a Competing NIH Grant Application Within the First 10 Years After Receiving the Doctorate for Selected NIH Grant Mechanisms, by Subgroup
Study Group
Average Award Rate for NIH Grant Applications Submitted in the First 10 Years After Receiving the Doctorate*
R01R01, R23, and
R29(Equivalents)
All RPGs Early K Awards
Late K Awards K Awards KRUMPT
# % # % # % # % # % # % # %Kirschstein-NRSA fellows 3,241 33.8 3,808 35.7 3,942 36.5 25 32.0 469 44.2 497 43.5 4.075 36.5Kirschstein-NRSA trainees 3,615 24.8 4,640 27.0 4,988 27.9 89 46.5 324 36.2 431 37.6 12,262 28.0Kirschstein-NRSA trainee/fellows
973 29.5 1,205 32.3 1,250 32.7 14 60.7 116 41.2 131 43.4 2,105 32.9
Unsuccessful fellowship applicants
2,896 28.8 3,357 29.9 3,590 30.9 74 38.3 335 39.9 426 40.3 9,516 31.0
Other postdoctorates 480 28.1 547 27.9 606 28.8 17 35.5 60 40.8 81 39.3 2,432 29.5Other research doctorates 517 19.2 557 19.4 674 21.2 23 34.1 36 32.3 68 46.5 5535 22.2* The average award rate is defined as the number of awards over the number of competing applications submitted in the first 10 years after receipt of the Ph.D. Source: CGAF
Table C.4. Mean Number of Years Between Receipt of the Ph.D. and First Competing Award for Selected Award Mechanisms, by Subgroup
Study Group
Mean Elapsed Number of Years From Ph.D. to Submission of First NIH Grant Award
R01R01, R23, and
R29(Equivalents)
All RPGs Early K Awards Late K Awards K Awards KRUMPT
# % # % # % # % # % # % # %Kirschstein-NRSA fellows 2,597 7.8 2,919 6.6 3,157 6.7 20 11.7 290 8.7 313 8.9 3,348 6.9Kirschstein-NRSA trainees 2,506 8.9 3,012 7.2 3,383 7.3 68 8.3 165 8.8 244 8.7 3.722 7.5Kirschstein-NRSA trainee/fellows
776 8.4 901 6.8 971 6.9 13 7.4 75 9.5 88 9.0 1.035 7.1
Unsuccessful fellowship applicants
2,102 8.0 2,354 7.0 2,620 7.1 52 8.4 177 8.3 238 8.0 2,886 7.3
Other postdoctorates 325 7.9 351 6.9 413 7.1 10 8.0 34 8.1 47 8.1 474 7.6Other research doctorates 257 7.4 278 6.8 371 6.6 13 7.4 19 6.8 36 7.3 458 6.7Source: CGAF
Table C.5. Number and Percentage of Kirschstein-NRSA Postdoctorates Submitting a Competing NIH Grant Application Within the First 10 Years After Receiving the Doctorate, by Subgroup
Study Group DegreeNumber of
Kirschstein-NRSA
Postdoctorates
Number and Percentage of Kirschstein-NRSA Postdoctorates
R01 R01, R23, and R29
RPGs Other Than R01,
R23, and R29All RPGs
Early Career
K Awards
Late Career
K AwardsAll K
Awards KRUMPT
# % # % # % # % # % # % # % # %
Kirschstein-NRSA fellows
M.D. 666 340 51.1 396 59.5 70 10.5 403 60.5 116 17.4 41 6.2 160 24.0 433 65.0M.D./Ph.D. 515 249 48.3 285 55.3 73 14.2 299 58.1 25 4.9 34 6.6 65 12.6 316 61.4Ph.D. 7,205 3,506 48.7 3,899 54.1 715 9.9 4,040 56.1 31 0.4 494 6.9 527 7.3 4,194 58.2Total 8,386 4,095 48.8 4,580 54.6 858 10.2 4,742 56.5 172 2.1 569 6.8 752 9.0 4,943 58.9
Kirschstein-NRSA trainees
M.D. 5,732 1,906 33.3 2,303 40.2 485 8.5 2,437 42.5 884 15.4 173 3.0 1,076 18.8 2,787 48.6M.D./Ph.D. 1,632 692 42.4 805 49.3 207 12.7 836 51.2 149 9.1 61 3.7 224 13.7 908 55.6Ph.D. 12,262 3,980 32.5 4,716 38.5 1,146 9.3 5,026 41.0 84 0.7 336 2.7 441 3.6 5,338 43.5Total 19,626 6,578 33.5 7,824 39.9 1,838 9.4 8,299 42.3 1,117 5.7 570 2.9 1,741 8.9 9,033 46.0
Kirschstein-NRSA trainee/fellows
M.D. 493 267 54.2 315 63.9 52 10.5 319 64.7 106 21.5 32 6.5 137 27.8 353 71.6M.D./Ph.D. 250 150 60.0 174 69.6 24 9.6 177 70.8 22 8.8 17 6.8 38 15.2 186 74.4Ph.D. 2,105 1,087 51.6 1,234 58.6 249 11.8 1,289 61.2 12 0.6 137 6.5 149 7.1 1,336 63.5Total 2,848 1,504 52.8 1,723 60.5 325 11.4 1,785 62.7 140 4.9 186 6.5 324 11.4 1,875 65.8
Total
M.D. 6,891 2,513 36.5 3,014 43.7 607 8.8 3,159 45.8 1,106 16.0 246 3.6 1,373 19.9 3,573 51.9M.D./Ph.D. 2,397 1,091 45.5 1,264 52.7 304 12.7 1,312 54.7 196 8.2 112 4.7 327 13.6 1,410 58.8Ph.D. 21,572 8,573 39.7 9,849 45.7 2,110 9.8 10,355 48.0 127 0.6 967 4.5 1,117 5.2 10,868 50.4Total 30,860 12,177 39.5 14,127 45.8 3021 9.8 14,826 48.0 1,429 4.6 1,325 4.3 2,817 9.1 15,851 51.4
Source: CGAF
Table C.6. Number and Percentage of Kirschstein-NRSA Postdoctorates Submitting a Competing NIH Grant Application Within the First 10 Years After Receiving the Doctorate Who Were Awarded a Grant for Selected NIH Grant Mechanisms, by Subgroup
Study Group DegreeNumber of
Kirschstein-NRSA
Postdoctorates
Number and Percentage of Kirschstein-NRSA Postdoctorates*
R01 R01, R23, and R29
RPGs Other Than R01,
R23, and R29All RPGs
Early Career
K AwardsLate Career K Awards
All K Awards KRUMPT
# % # % # % # % # % # % # % # %
Kirschstein-NRSA fellows
M.D. 666 188 55.3 229 57.8 34 48.6 240 59.6 73 62.9 20 48.8 96 60.0 279 64.4M.D./Ph.D. 515 157 63.1 199 69.8 38 52.1 213 71.2 15 60.0 20 58.8 39 60.0 233 73.7Ph.D. 7,205 2,219 63.3 2,631 67.5 388 54.3 2,814 69.7 14 45.2 246 49.8 261 49.5 2,951 70.4Total 8,386 2,564 62.6 3,059 66.8 460 53.6 3,267 68.9 102 59.3 286 50.3 396 52.7 3,463 70.1
Kirschstein-NRSA trainees
M.D. 5,732 957 50.2 1,253 54.4 218 44.9 1,359 55.8 535 60.5 80 46.2 641 59.6 1,719 61.7M.D./Ph.D. 1,632 396 57.2 489 60.7 98 47.3 525 62.8 86 57.7 34 55.7 137 61.2 600 66.1Ph.D. 12,262 1,990 50.0 2,596 55.0 523 45.6 2,885 57.4 42 50.0 142 42.3 189 42.9 3,113 58.3Total 19,626 3,343 50.8 4,338 55.4 839 45.6 4,769 57.5 663 59.4 256 44.9 967 55.5 5,432 60.1
Kirschstein-NRSA trainee/fellows
M.D. 493 152 56.9 192 61.0 19 36.5 201 63.0 72 67.9 18 56.3 90 65.7 246 69.7M.D./Ph.D. 250 90 60.0 114 65.5 14 58.3 120 67.8 12 54.5 6 35.3 18 47.4 132 71.0Ph.D. 2,105 672 61.8 834 67.6 116 46.6 879 68.2 10 83.3 64 46.7 74 49.7 929 69.5Total 2,848 914 60.8 1,140 66.2 149 45.8 1,200 67.2 94 67.1 88 47.3 182 56.2 1,307 69.7
Total
M.D. 6,891 1,297 51.6 1,674 55.5 271 44.6 1,800 57.0 680 61.5 118 48.0 827 60.2 2,244 62.8M.D./Ph.D. 2,397 643 58.9 802 63.4 150 49.3 858 65.4 113 57.7 60 53.6 194 59.3 965 68.4Ph.D. 21,572 4,881 56.9 6,061 61.5 1,027 48.7 6,578 63.5 66 52.0 452 46.7 524 46.9 6,993 64.3Total 30,860 6,821 56.0 8,537 60.4 1,448 47.9 9,236 62.3 859 60.1 630 47.5 1,545 54.8 10,202 64.4
* The denominators for the percentages are provided in table C.5 as cell observations.Source: CGAF
Table C.7. Average Award Rate* of Those Submitting a Competing NIH Grant Application Within the First 10 Years After Receiving the Doctorate for Selected NIH Grant Mechanisms, by Subgroup
Study Group Degree
Number and Percentage of Kirschstein-NRSA Postdoctorates
R01 R01, R23, and R29
RPGs Other Than R01,
R23, and R29All RPGs Early Career
K AwardsLate Career K Awards All K Awards KRUMPT
# % # % # % # % # % # % # % # %
Kirschstein-NRSA fellows
M.D. 340 25.0 396 26.2 70 40.4 403 27.1 116 56.0 41 43.3 160 51.2 433 30.0M.D./Ph.D. 249 30.1 285 31.6 73 39.2 299 32.1 25 58.0 34 57.4 65 57.7 316 34.7Ph.D. 3,506 31.4 3,899 32.6 715 44.3 4,040 33.6 31 41.9 494 45.8 527 45.4 4,194 34.0Total 4,095 30.8 4,580 32.0 858 43.5 4,742 33.0 172 53.7 569 46.3 752 47.7 4,943 33.7
Kirschstein-NRSA trainees
M.D. 1,906 24.7 2,303 26.0 485 35.5 2,437 26.7 884 50.4 173 43.3 1,076 49.3 2,787 30.4M.D./Ph.D. 692 27.6 805 28.7 207 35.2 836 29.1 149 52.0 61 52.5 224 54.5 908 31.4Ph.D. 3,980 24.0 4,716 25.7 1,146 34.6 5,026 26.7 84 41.9 336 38.0 441 37.9 5,338 27.1Total 6,578 24.6 7,824 26.1 1,838 34.9 8,299 26.9 1,117 50.0 570 41.1 1,741 47.1 9,033 28.5
Kirschstein-NRSA trainee/fellows
M.D. 267 26.6 315 28.3 52 28.2 319 28.8 106 58.8 32 49.0 137 54.5 353 34.2M.D./Ph.D. 150 29.3 174 31.1 24 51.4 177 31.7 22 52.3 17 32.4 38 44.3 186 33.0Ph.D. 1,087 28.1 1,234 30.2 249 35.8 1,289 30.5 12 62.5 137 43.7 149 45.1 1,336 30.9Total 1,504 28.0 1,723 30.0 325 35.8 1,785 30.3 140 58.1 186 43.5 324 49.0 1,875 31.8
Total
M.D. 2,513 24.9 3,014 26.3 607 35.4 3,159 27.0 1,106 51.8 246 44.0 1,373 50.0 3,573 30.8M.D./Ph.D. 1,091 28.4 1,264 29.7 304 37.4 1,312 30.1 196 52.8 112 50.9 327 54.0 1,410 32.3Ph.D. 8,573 27.5 9,849 29.0 2,110 38.0 10,355 29.9 127 43.8 967 42.8 1,117 42.4 10,868 30.2Total 12,177 27.1 14,127 28.5 3,021 37.4 14,826 29.3 1,429 51.2 1,325 43.7 2,817 47.5 15,851 30.5
* The average award rate is defined as the number of awards over the number of competing applications submitted in the first 10 years after receipt of the Ph.D. Source: CGAF
Table C.8. Mean Elapsed Number of Years Between Last Year of Kirschstein-NRSA Support and First NIH Competing Award for Selected NIH Grant Mechanisms, by Subgroup
Study Group Degree
Number of Kirschstein-NRSA Postdoctorates (Includes Only Those Receiving an Award)
R01 R01, R23, and R29
RPGs Other Than R01,
R23, and R29All RPGs Early Career
K AwardsLate Career K Awards All K Awards KRUMPT
# Years # Years # Years # Years # Years # Years # Years # Years
Kirschstein-NRSA fellows
M.D. 233 6.8 263 5.2 84 11.5 279 5.2 98 1.9 30 6.7 133 3.0 325 3.9M.D./Ph.D. 184 5.5 221 4.1 86 10.4 240 4.2 18 2.0 24 5.0 51 4.9 260 3.9Ph.D. 2,597 5.3 2,919 4.1 885 10.6 3,157 4.1 20 9.4 290 6.4 313 6.5 3,348 4.3Total 3,014 5.4 3,403 4.2 1,055 10.6 3,676 4.2 136 3.0 344 6.3 497 5.4 3,933 4.3
Kirschstein-NRSA trainees
M.D. 1,263 7.6 1,480 6.1 529 11.5 1,628 6.3 774 2.0 131 9.2 928 3.0 2,073 4.6M.D./Ph.D. 488 6.4 557 5.0 203 10.3 608 5.2 142 2.7 48 6.8 209 3.6 705 4.3Ph.D. 2,506 6.5 3,012 4.8 1,021 9.4 3,383 4.8 68 5.5 165 6.5 244 6.2 3,722 5.0Total 4,257 6.8 5,049 5.2 1,753 10.2 5,619 5.3 984 2.3 344 7.6 1,381 3.6 6,500 4.8
Kirschstein-NRSA trainee/fellows
M.D. 171 6.2 206 4.6 60 11.6 215 4.5 132 1.1 24 8.5 152 2.0 282 2.5M.D./Ph.D. 102 5.4 123 3.5 33 9.7 129 3.6 21 2.3 9 7.3 29 3.6 142 3.1Ph.D. 776 5.3 901 3.7 257 10.1 971 3.9 13 2.5 75 6.7 88 6.0 1,035 4.0Total 1,049 5.5 1,230 3.9 350 10.3 1,315 4.0 166 1.4 108 7.1 269 3.5 1,459 3.7
Total
M.D. 1,667 7.4 1,949 5.8 673 11.5 2,122 6.0 1,004 1.9 185 8.7 1,213 2.8 2,680 4.3M.D./Ph.D. 774 6.1 901 4.6 322 10.2 977 4.7 181 2.6 81 6.3 289 3.8 1,107 4.1Ph.D. 5,879 5.8 6,832 4.4 2,163 10.0 7,511 4.4 101 5.9 530 6.4 645 6.3 8,105 4.6Total 8,320 6.2 9,682 4.7 3,158 10.3 10,610 4.8 1,286 2.3 796 7.0 2,147 4.0 11,892 4.5
PUBLICATION OUTCOMES
The data represent a information on publications by a subsample of research doctorates drawn from the NIH predoctoral study (See Pion for details). In broad terms, that study furnished Thompson ISI with identifiers of a sample of predoctoral students for extracting publication records in their database. These individuals were then matched to the individuals identified for this study, and this matched subsample was used to analyze publication patterns among the subgroups.
Table C.9. Percentage of Research Doctorates Who Have Published Between Receipt of the Ph.D. and 1995, by Ph.D. Cohort, Number of Publications, and Study Group
Study Group
Published At Least Once Between Receipt of Ph.D. and 1995
More than 10 Publications Between Receipt of Ph.D. and 1995
More than 20 Publications Between Receipt of Ph.D. and 1995
1981–82 Ph.D.’s 1987–88 Ph.D.’s 1981–82 Ph.D.’s 1987–88 Ph.D.’s 1981–82 Ph.D.’s 1987–88 Ph.D.’s# % Total # % Total # % Total # % Total # % Total # % Total
Kirschstein-NRSA fellows
119 86.2 138 93 81.6 114 67 48.6 138 22 19.3 114 30 21.7 138 1 0.9 114
Kirschstein-NRSA trainees
251 83.7 300 187 77.6 241 131 43.7 300 47 19.5 241 52 17.3 300 14 5.8 241
Kirschstein-NRSA trainee/fellows
45 91.8 49 37 90.2 41 25 51.0 49 11 26.8 41 9 18.4 49 2 4.9 41
Unsuccessful fellowship applicants
129 83.2 155 119 81.0 147 54 34.8 155 27 18.4 147 24 15.5 155 6 4.1 147
Other postdoctorates
159 82.4 193 178 80.2 222 63 32.6 193 53 23.9 222 22 11.4 193 4 1.8 222
Other research doctorates
283 64.0 442 287 60.4 475 61 13.8 442 53 11.2 475 25 5.7 442 12 1.0 475
Source: ISI
Table C.10. Median and Mean Number of Publications Between Receipt of the Ph.D and 1995,by Ph.D. Cohort and Study Group
Study Group1981–82 Ph.D.’s 1987–88 Ph.D.’s
# Median Mean # Median MeanKirschstein-NRSA fellows 138 10.0 14.4 114 4.5 5.9Kirschstein-NRSA trainees 300 9.0 11.9 241 5.0 7.0Kirschstein-NRSA trainee/fellows 49 11.0 12.6 41 6.0 7.8Unsuccessful fellowship applicants 155 8.0 10.2 147 5.0 6.8Other postdoctorates 193 5.0 9.4 222 5.0 6.4Other research doctorates 442 1.0 4.9 475 1.0 4.7Source: ISI
Table C.11. Median and Mean Elapsed Years to First Publication After Receipt of the Ph.D, by Ph.D. Cohort and Study Group
Study Group1981–82 Ph.D.’s 1987–88 Ph.D.’s
# Median Mean # Median MeanKirschstein-NRSA fellows 119 1.0 2.1 92 1.0 1.7Kirschstein-NRSA trainees 246 1.0 1.9 183 1.0 1.4Kirschstein-NRSA trainee/fellows 44 1.0 1.5 37 1.0 1.0Unsuccessful fellowship applicants 128 1.0 2.2 117 1.0 1.5Other postdoctorates 157 1.0 2.4 175 1.0 1.8Other research doctorates 276 2.0 3.0 277 1.0 1.8Source: ISI
Table C.12. Median and Mean Number of Average Citations Per Publication Between Receipt of the Ph.D. and 1995, By Ph.D. Cohort and Study Group
Study Group1981–82 Ph.D.’s 1987–88 Ph.D.’s
# Median Mean # Median MeanKirschstein-NRSA fellows 63 22.3 36.1 56 17.6 25.5Kirschstein-NRSA trainees 128 21.1 27.8 94 13.2 21.9Kirschstein-NRSA trainee/fellows 22 26.1 30.6 19 15.3 26.0Unsuccessful fellowship applicants 75 23.7 29.2 67 19.0 26.2Other postdoctorates 82 16.1 18.0 84 12.3 21.1Other research doctorates 146 8.8 19.5 153 8.6 14.1Source: ISI
EMPLOYMENT OUTCOMES
Employment data derive directly from the SDR and pertain only to those individuals who received a Ph.D. The percentages in the following two tables are calculated from weighted observations, while the number of observations reported are unweighted. The weights were those reported on the 1995 SDR.
Table C.13. Number and Percentage of Ph.D.’s, by Employment Type, Years From Receipt of Research Doctorate, and Study Group
Study GroupYears From Receipt of Research Doctorate
Employment TypeTotal
Business/Industry University or Affiliated Research Facility Other*
# % # % # % #Kirschstein-NRSA fellows
4 years 31 10.8 185 64.5 66 24.7 2828 years 71 15.8 308 64.8 98 19.4 477
12 years 111 26.0 265 57.2 83 16.9 459Kirschstein-NRSA trainees
4 years 38 8.8 265 58.8 144 32.4 4478 years 123 19.1 371 57.2 150 23.7 644
12 years 148 22.5 359 54.1 137 23.3 644Kirschstein-NRSA trainee/fellows
4 years 6 8.6 46 64.9 17 26.5 698 years 27 17.9 90 60.5 27 21.6 144
12 years 32 22.8 87 56.2 27 21.1 146Unsuccessful fellowship applicants
4 years 63 22.1 152 46.2 96 31.8 3118 years 92 22.0 251 53.8 108 24.2 451
12 years 109 24.7 230 53.2 87 22.0 426Other postdoctorates
4 years 182 13.5 650 45.4 532 41.2 1,3648 years 279 15.5 846 46.4 629 38.2 1,754
12 years 280 16.3 770 45.4 558 38.3 1,608Other research doctorates
4 years 352 8.8 1,739 43.6 1,686 47.7 3,7778 years 505 11.6 1,838 41.8 1,847 46.6 4,190
12 years 559 13.6 1,679 39.9 1,769 46.5 4,007* Other includes self-employed; primary or secondary school; U.S., State, or local government; nonacademic hospital (clinic); U.S. military; nonprofit organization; unknown; and other.
Table C.14. Number and Percentage of Ph.D.’s, by Academic Rank and Tenure Status, Years From Receipt of Research Doctorate, and Study Group
Study GroupYears From Receipt of Research Doctorate
Academic Rank Percentage Who AchieveTenureAssistant
ProfessorAssociate Professor Professor Other Total
# % # % # % # % # % # TotalKirschstein-NRSA fellows
4 years 125 68.9 2 2.7 0 0.0 49 28.5 176 0.6 1 1068 years 217 71.6 41 16.1 0 0.0 32 12.3 290 12.9 25 216
12 years 75 27.7 129 63.0 12 5.4 11 3.9 227 55.1 104 201Kirschstein-NRSA trainees
4 years 155 66.2 5 3.2 0 0.0 84 30.7 244 3.0 3 1148 years 226 66.2 52 17.0 2 0.8 53 16.0 333 18.0 30 201
12 years 116 36.0 128 46.5 14 6.2 37 11.3 295 49.0 100 225Kirschstein-NRSA trainee/fellows
4 years 22 56.3 3 5.7 0 0.0 18 38.1 43 4.7 1 188 years 65 76.5 8 12.0 0 0.0 9 11.5 82 6.6 3 57
12 years 28 36.5 36 49.7 3 6.8 5 7.0 72 55.3 30 58Unsuccessful fellowship applicants
4 years 100 68.6 3 1.6 1 0.6 36 29.3 140 0.0 0 898 years 157 70.1 33 13.6 5 2.7 28 13.7 223 16.4 30 173
12 years 55 27.9 90 49.0 10 5.8 26 17.3 181 53.5 85 161Other postdoctorates
4 years 422 74.6 31 4.9 1 0.1 121 20.4 575 7.5 26 3828 years 403 51.2 221 34.5 13 1.3 98 13.0 735 40.3 193 561
12 years 136 18.1 340 53.8 103 17.8 69 10.4 648 70.8 382 564Other research doctorates
4 years 1,076 67.3 286 16.4 41 2.6 201 13.7 1,604 18.5 232 1,2198 years 478 28.6 851 52.8 157 9.9 139 8.7 1,625 64.7 864 1,368
12 years 143 9.0 745 50.5 435 30.9 149 9.6 1,472 78.7 1,011 1,301
APPENDIX D
REGRESSION TABLES
Table D.1. Regression Results Reflecting the Likelihood of Individuals Submitting a Competing R01 Application
Model Information
Dataset WORK.PHD_GRANT_POSTDOC
Distribution Binomial
Link function Logit
Dependent variable Ph.D.’s who applied for R01 awards
Observations
Number of observations read 39,055
Number of observations used 39,045
Number of events 14,357
Number of trials 39,045
Missing values 10
Response Profile
OrderedValue allanyapp01 Total
Frequency
1 1 14,357
2 0 24,688
PROC GENMOD is modeling the probability that Ph.D.’s who applied for R01 awards = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 39E3 46643.9825 1.1952
Scaled deviance 39E3 46643.9825 1.1952
Pearson Chi-Square 39E3 38877.6455 0.9962
Scaled Pearson X2 39E3 38877.6455 0.9962
Log likelihood -23321.9913
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 2.5908 0.1974 2.2038 2.9778 172.18 <.0001
Ph.D. year 1 -0.0492 0.0022 -0.0535 -0.0449 508.99 <.0001
Study group
Kirschstein-NRSA fellows 1 1.9279 0.0579 1.8144 2.0413 1109.81 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.6656 0.0691 0.5302 0.8010 92.82 <.0001
Kirschstein-NRSA trainees 1 1.3845 0.0544 1.2780 1.4911 648.20 <.0001
Kirschstein-NRSA trainee/fellows
1 2.0171 0.0692 1.8815 2.1528 849.73 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 1.3816 0.0552 1.2734 1.4897 626.54 <.0001
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2552 0.0240 -0.3022 -0.2083 113.48 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 -0.0952 0.0648 -0.2221 0.0317 2.16 0.1415
Postdoctoral training 1 0.3197 0.0576 0.2068 0.4327 30.79 <.0001
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.1738 0.0275 -0.2277 -0.1198 39.84 <.0001
9+ 1 -0.6093 0.0466 -0.7006 -0.5179 170.93 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0659 0.0113 0.0437 0.0880 33.97 <.0001
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.0779 0.0295 -0.1357 -0.0201 6.97 0.0083
Ranked from 26 to 100 1 0.0266 0.0279 -0.0281 0.0813 0.91 0.3409
Ranked from 11 to 25 1 0.0849 0.0392 0.0080 0.1618 4.68 0.0305
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 -0.4342 0.0813 -0.5936 -0.2748 28.50 <.0001
Behavioral 1 -0.5011 0.0683 -0.6350 -0.3672 53.79 <.0001
Biological 1 -0.3554 0.0666 -0.4858 -0.2249 28.51 <.0001
Chemistry 1 -1.2865 0.0767 -1.4367 -1.1363 281.69 <.0001
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 302.59 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees 1 0.01 0.9206 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 194.37 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 3.08 0.0792 LR
All Kirschstein-NRSA postdoctorates versus other study groups 1 40.78 <.0001 LR
Table D.2. Regression Results Reflecting the Likelihood of Individuals Submitting a Competing R01-Equivalent Application
Model Information
Dataset WORK.PHD_GRANT_POSTDOC
Distribution Binomial
Link function Logit
Dependent variable Ph.D.’s who applied for R01-equivalent awards
Observations
Number of observations read 39,055
Number of observations used 39,045
Number of events 15,825
Number of trials 39,045
Missing values 10
Response Profile
OrderedValue totrnewapall Total
Frequency
1 1 15,825
2 0 23,220
PROC GENMOD is modeling the probability that Ph.D.’s who applied for R01-equivalent awards = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 39E3 47425.8072 1.2153
Scaled deviance 39E3 47425.8072 1.2153
Pearson Chi-Square 39E3 38828.4134 0.9950
Scaled Pearson X2 39E3 38828.4134 0.9950
Log likelihood -23712.9036
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 2.1096 0.1946 1.7282 2.4909 117.56 <.0001
Ph.D. year 1 -0.0419 0.0021 -0.0461 -0.0377 381.72 <.0001
Study group
Kirschstein-NRSA fellows 1 2.0101 0.0570 1.8983 2.1218 1243.68 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.6726 0.0675 0.5403 0.8050 99.19 <.0001
Kirschstein-NRSA trainees 1 1.4968 0.0532 1.3925 1.6011 790.53 <.0001
Kirschstein-NRSA trainee/fellows
1 2.1459 0.0690 2.0107 2.2811 967.66 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 1.4168 0.0541 1.3107 1.5228 685.56 <.0001
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2437 0.0236 -0.2900 -0.1974 106.43 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 -0.1285 0.0637 -0.2533 -0.0037 4.07 0.0436
Postdoctoral training 1 0.3244 0.0568 0.2132 0.4357 32.66 <.0001
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.1894 0.0274 -0.2431 -0.1357 47.83 <.0001
9+ 1 -0.5897 0.0453 -0.6784 -0.5010 169.77 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0676 0.0113 0.0454 0.0897 35.77 <.0001
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.0382 0.0291 -0.0952 0.0189 1.72 0.1898
Ranked from 26 to 100 1 0.0437 0.0277 -0.0105 0.0980 2.49 0.1142
Ranked from 11 to 25 1 0.1049 0.0389 0.0287 0.1810 7.29 0.0070
Top 10 institution 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 -0.4434 0.0800 -0.6003 -0.2865 30.69 <.0001
Behavioral 1 -0.5400 0.0672 -0.6718 -0.4082 64.51 <.0001
Biological 1 -0.3640 0.0656 -0.4925 -0.2354 30.79 <.0001
Chemistry 1 -1.3613 0.0754 -1.5091 -1.2136 326.01 <.0001
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 271.33 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees 1 7.39 0.0065 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 269.68 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 6.90 0.0086 LR
Kirschstein-NRSA postdoctorates versus other study groups 1 46.99 <.0001 LR
Table D.3. Regression Results Reflecting the Likelihood of a Competing R01 Award Given the Submission of an Application
Model Information
Dataset WORK.PHD_GRANT_POSTDOC1
Distribution Binomial
Link function Logit
Dependent variable Ph.D.’s awarded R01 awards
Observations
Number of observations read 14,362
Number of observations used 14,357
Number of events 8,537
Number of trials 14,357
Missing values 5
Response Profile
OrderedValue allanyawd01 Total
Frequency
1 1 8,537
2 0 5,820
PROC GENMOD is modeling the probability that Ph.D.’s awarded R01 awards = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 14E3 18733.7673 1.3067
Scaled deviance 14E3 18733.7673 1.3067
Pearson Chi-Square 14E3 14360.1412 1.0016
Scaled Pearson X2 14E3 14360.1412 1.0016
Log likelihood -9366.8836
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 1.0675 0.3254 0.4296 1.7053 10.76 0.0010
Ph.D. year 1 -0.0117 0.0036 -0.0187 -0.0046 10.61 0.0011
Study group
Kirschstein-NRSA fellows 1 0.6429 0.1020 0.4429 0.8429 39.70 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.2114 0.1254 -0.0343 0.4571 2.84 0.0917
Kirschstein-NRSA trainees 1 0.1426 0.0987 -0.0508 0.3360 2.09 0.1484
Kirschstein-NRSA trainee/fellows
1 0.5180 0.1142 0.2941 0.7418 20.57 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 0.3770 0.0998 0.1813 0.5727 14.26 0.0002
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2131 0.0379 -0.2874 -0.1388 31.58 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 -0.4712 0.1044 -0.6759 -0.2666 20.37 <.0001
Postdoctoral training 1 0.1067 0.0914 -0.0726 0.2859 1.36 0.2435
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.1179 0.0425 -0.2011 -0.0346 7.70 0.0055
9+ 1 -0.6639 0.0786 -0.8179 -0.5098 71.37 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0779 0.0174 0.0438 0.1120 20.01 <.0001
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.1565 0.0463 -0.2473 -0.0658 11.43 0.0007
Ranked from 26 to 100 1 0.0737 0.0435 -0.0117 0.1590 2.86 0.0907
Ranked from 11 to 25 1 0.0225 0.0610 -0.0971 0.1420 0.14 0.7125
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 -0.1494 0.1298 -0.4039 0.1051 1.32 0.2498
Behavioral 1 -0.0636 0.1115 -0.2821 0.1550 0.33 0.5686
Biological 1 0.0998 0.1078 -0.1115 0.3110 0.86 0.3546
Chemistry 1 0.0166 0.1272 -0.2327 0.2660 0.02 0.8961
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 115.38 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees 1 24.38 <.0001 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates 1 0.56 0.4562 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 3.07 0.0797 LR
All Kirschstein-NRSA postdoctorates versus other study groups 1 0.46 0.4998 LR
Table D.4. Regression Results Reflecting the Likelihood of a Competing R01-Equivalent Award Given the Submission of an Application
Model Information
Dataset WORK.PHD_GRANT_POSTDOC1
Distribution Binomial
Link function Logit
Dependent variable Ph.D.’s awarded R01-equivalent awards
Observations
Number of observations read 15,830
Number of observations used 15,825
Number of events 9,786
Number of trials 15,825
Missing values 5
Response Profile
OrderedValue totrnewawall Total
Frequency
1 1 9,786
2 0 6,039
PROC GENMOD is modeling the probability that Ph.D.’s awarded R01-equivalent awards = ”1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 16E3 20336.9781 1.2867
Scaled deviance 16E3 20336.9781 1.2867
Pearson Chi-Square 16E3 15821.8194 1.0011
Scaled Pearson X2 16E3 15821.8194 1.0011
Log likelihood -10168.4891
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 0.4087 0.3136 -0.2059 1.0232 1.70 0.1925
Ph.D. year 1 -0.0042 0.0034 -0.0110 0.0025 1.50 0.2211
Study group
Kirschstein-NRSA fellows 1 0.6991 0.0989 0.5053 0.8929 50.00 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.1601 0.1212 -0.0776 0.3977 1.74 0.1867
Kirschstein-NRSA trainees 1 0.1941 0.0953 0.0073 0.3809 4.15 0.0417
Kirschstein-NRSA trainee/fellows
1 0.6021 0.1108 0.3849 0.8193 29.51 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 0.3826 0.0967 0.1932 0.5721 15.67 <.0001
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2093 0.0363 -0.2803 -0.1382 33.29 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 -0.3263 0.0995 -0.5213 -0.1314 10.76 0.0010
Postdoctoral training 1 0.1924 0.0868 0.0222 0.3626 4.91 0.0267
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.0803 0.0410 -0.1607 -0.0000 3.84 0.0500
9+ 1 -0.6619 0.0732 -0.8054 -0.5184 81.72 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0711 0.0167 0.0385 0.1037 18.23 <.0001
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.1657 0.0445 -0.2528 -0.0785 13.88 0.0002
Ranked from 26 to 100 1 0.0224 0.0421 -0.0601 0.1048 0.28 0.5947
Ranked from 11 to 25 1 -0.0104 0.0588 -0.1256 0.1047 0.03 0.8590
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 -0.0879 0.1233 -0.3295 0.1538 0.51 0.4760
Behavioral 1 -0.0710 0.1061 -0.2789 0.1369 0.45 0.5031
Biological 1 0.1438 0.1026 -0.0573 0.3449 1.96 0.1610
Chemistry 1 0.0553 0.1219 -0.1836 0.2942 0.21 0.6501
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 126.64 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees 1 17.25 <.0001 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 0.15 0.7020 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 1.94 0.1639 LR
Kirschstein-NRSA postdoctorates versus other study groups 1 0.22 0.6395 LR
Table D.5. Regression Results Relating to the Number of Competing R01 Awards
Model Information
Dataset WORK.PHD_GRANT_POSTDOC1
Distribution Poisson
Link function Log
Dependent variable Number of R01 awards received by Ph.D.’s
Observations
Number of observations read 14,362
Number of observations used 14,357
Missing values 5
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 14E3 38197.1905 2.6642
Scaled deviance 14E3 38197.1905 2.6642
Pearson Chi-Square 14E3 40187.5867 2.8031
Scaled Pearson X2 14E3 40187.5867 2.8031
Log likelihood -8237.5038
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 3.9367 0.1206 3.7004 4.1730 1066.33 <.0001
Ph.D. year 1 -0.0440 0.0013 -0.0466 -0.0415 1139.00 <.0001
Study group
Kirschstein-NRSA fellows 1 0.4493 0.0440 0.3631 0.5355 104.43 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.1703 0.0514 0.0695 0.2711 10.97 0.0009
Kirschstein-NRSA trainees 1 0.1345 0.0438 0.0485 0.2204 9.41 0.0022
Kirschstein-NRSA trainee/fellows
1 0.3801 0.0473 0.2874 0.4729 64.53 <.0001
Unsuccessful fellowship applicants
1 0.3281 0.0437 0.2424 0.4138 56.31 <.0001
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.1742 0.0142 -0.2020 -0.1465 151.39 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Expressed postdoctoral plans
Employment 1 -0.3705 0.0389 -0.4468 -0.2943 90.84 <.0001
Postdoctoral training 1 0.1501 0.0314 0.0885 0.2116 22.82 <.0001
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.1548 0.0141 -0.1825 -0.1272 120.42 <.0001
9+ 1 -0.6972 0.0377 -0.7711 -0.6232 341.66 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0351 0.0060 0.0233 0.0469 34.11 <.0001
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.1142 0.0169 -0.1474 -0.0810 45.46 <.0001
Ranked from 26 to 100 1 0.0528 0.0149 0.0237 0.0820 12.60 0.0004
Ranked from 11 to 25 1 -0.0005 0.0218 -0.0431 0.0422 0.00 0.9819
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 -0.1497 0.0549 -0.2572 -0.0421 7.44 0.0064
Behavioral 1 -0.1017 0.0477 -0.1952 -0.0081 4.53 0.0332
Biological 1 0.0438 0.0456 -0.0456 0.1333 0.92 0.3368
Chemistry 1 0.0827 0.0505 -0.0163 0.1817 2.68 0.1015
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 376.41 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees
1 120.72 <.0001 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 1.13 0.2879 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 9.20 0.0024 LR
All Kirschstein-NRSA postdoctorates versus other study groups 1 0.61 0.4355 LR
Table D.6. Regression Results Relating to the Number of Competing R01-Equivalent Awards
Model Information
Dataset WORK.PHD_GRANT_POSTDOC1
Distribution Poisson
Link function Log
Dependent variable Number of R01-equivalent awards received by Ph.D.’s
Observations
Number of observations read 15,830
Number of observations used 15,825
Missing values 5
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 16E3 41019.4016 2.5953
Scaled deviance 16E3 41019.4016 2.5953
Pearson Chi-Square 16E3 43001.6194 2.7208
Scaled Pearson X2 16E3 43001.6194 2.7208
Log likelihood -8971.5388
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 3.5224 0.1144 3.2981 3.7467 947.49 <.0001
Ph.D. year 1 -0.0387 0.0012 -0.0411 -0.0363 991.03 <.0001
Study group
Kirschstein-NRSA fellows 1 0.4431 0.0425 0.3598 0.5265 108.53 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.1679 0.0499 0.0702 0.2657 11.34 0.0008
Kirschstein-NRSA trainees 1 0.1304 0.0424 0.0473 0.2134 9.47 0.0021
Kirschstein-NRSA trainee/fellows
1 0.3874 0.0455 0.2981 0.4767 72.32 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 0.3080 0.0424 0.2250 0.3910 52.87 <.0001
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.1691 0.0133 -0.1951 -0.1431 162.63 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Expressed postdoctoral plans
Employment 1 -0.3308 0.0372 -0.4036 -0.2580 79.23 <.0001
Postdoctoral training 1 0.1617 0.0300 0.1030 0.2205 29.09 <.0001
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.1387 0.0134 -0.1651 -0.1124 106.75 <.0001
9+ 1 -0.6735 0.0343 -0.7406 -0.6064 386.55 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0344 0.0055 0.0236 0.0453 38.57 <.0001
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.1208 0.0159 -0.1520 -0.0896 57.51 <.0001
Ranked from 26 to 100 1 0.0444 0.0141 0.0168 0.0720 9.91 0.0016
Ranked from 11 to 25 1 -0.0083 0.0205 -0.0486 0.0319 0.16 0.6848
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 -0.1386 0.0513 -0.2391 -0.0381 7.31 0.0069
Behavioral 1 -0.1014 0.0447 -0.1890 -0.0137 5.13 0.0235
Biological 1 0.0477 0.0427 -0.0360 0.1314 1.25 0.2642
Chemistry 1 0.0841 0.0475 -0.0089 0.1771 3.14 0.0764
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 422.51 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees 1 114.51 <.0001 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 1.35 0.2461 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 6.78 0.0092 LR
Kirschstein-NRSA postdoctorates versus other study groups 1 0.66 0.4152 LR
Table D.7. Regression Results Relating to the Individual Average Award Rate for Competing R01 Awards
Model Information
Dataset WORK.PHD_GRANT_POSTDOC1
Distribution Normal
Link function Identity
Dependent variable Award rate for Ph.D.’s receiving R01 awards
Observations
Number of observations read 14,362
Number of observations used 14,357
Missing values 5
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 14E3 1046.8238 0.0730
Scaled deviance 14E3 14357.0000 1.0014
Pearson Chi-Square 14E3 1046.8238 0.0730
Scaled Pearson X2 14E3 14357.0000 1.0014
Log likelihood -1574.9635
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 0.1919 0.0424 0.1087 0.2751 20.44 <.0001
Ph.D. year 1 0.0005 0.0005 -0.0005 0.0014 0.98 0.3219
Study group
Kirschstein-NRSA fellows 1 0.0567 0.0133 0.0307 0.0828 18.18 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.0261 0.0165 -0.0061 0.0584 2.53 0.1120
Kirschstein-NRSA trainees 1 -0.0066 0.0129 -0.0319 0.0186 0.26 0.6068
Kirschstein-NRSA trainee/fellows
1 0.0309 0.0148 0.0018 0.0600 4.33 0.0374
Unsuccessful Kirschstein-NRSA fellowship applicants
1 0.0353 0.0131 0.0097 0.0609 7.29 0.0069
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.0173 0.0050 -0.0270 -0.0076 12.21 0.0005
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Expressed postdoctoral plans
Employment 1 -0.0580 0.0137 -0.0848 -0.0312 18.02 <.0001
Postdoctoral training 1 0.0019 0.0119 -0.0213 0.0252 0.03 0.8699
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.0183 0.0055 -0.0290 -0.0075 11.10 0.0009
9+ 1 -0.0791 0.0102 -0.0992 -0.0590 59.65 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0127 0.0022 0.0084 0.0170 33.39 <.0001
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.0245 0.0060 -0.0363 -0.0127 16.44 <.0001
Ranked from 26 to 100 1 0.0143 0.0056 0.0033 0.0253 6.48 0.0109
Ranked from 11 to 25 1 0.0055 0.0079 -0.0100 0.0209 0.48 0.4898
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 -0.0027 0.0170 -0.0361 0.0306 0.03 0.8738
Behavioral 1 0.0059 0.0146 -0.0227 0.0345 0.16 0.6867
Biological 1 0.0157 0.0141 -0.0120 0.0433 1.23 0.2667
Chemistry 1 0.0262 0.0166 -0.0063 0.0588 2.50 0.1140
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 1 0.2700 0.0016 0.2669 0.2732
Note: The scale parameter was estimated by maximum likelihood.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 110.45 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees 1 44.90 <.0001 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 7.23 0.0072 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 8.21 0.0042 LR
All Kirschstein-NRSA postdoctorates versus other study groups 1 0.26 0.6097 LR
Table D.8. Regression Results Relating to the Individual Average Award Rate for Competing R01-Equivalent Awards
Model Information
Dataset WORK.PHD_GRANT_POSTDOC1
Distribution Normal
Link function Identity
Dependent variable Award rate for Ph.D.’s receiving R01-equivalent awards
Observations
Number of observations read 15,830
Number of observations used 15,825
Missing values 5
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 16E3 1149.5285 0.0727
Scaled deviance 16E3 15825.0000 1.0013
Pearson Chi-Square 16E3 1149.5285 0.0727
Scaled Pearson X2 16E3 15825.0000 1.0013
Log likelihood -1706.2352
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 0.1915 0.0406 0.1119 0.2711 22.25 <.0001
Ph.D. year 1 0.0003 0.0004 -0.0006 0.0011 0.33 0.5635
Study group
Kirschstein-NRSA fellows 1 0.0667 0.0129 0.0414 0.0920 26.69 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.0214 0.0160 -0.0099 0.0526 1.79 0.1808
Kirschstein-NRSA trainees 1 0.0061 0.0125 -0.0185 0.0306 0.24 0.6272
Kirschstein-NRSA trainee/fellows
1 0.0437 0.0143 0.0156 0.0718 9.27 0.0023
Unsuccessful Kirschstein-NRSA fellowship applicants
1 0.0405 0.0127 0.0156 0.0654 10.18 0.0014
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.0161 0.0047 -0.0252 -0.0069 11.76 0.0006
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Expressed postdoctoral plans
Employment 1 -0.0426 0.0130 -0.0681 -0.0171 10.74 0.0010
Postdoctoral training 1 0.0109 0.0113 -0.0112 0.0330 0.94 0.3333
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.0149 0.0052 -0.0251 -0.0046 8.09 0.0045
9+ 1 -0.0775 0.0095 -0.0962 -0.0588 65.84 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0109 0.0021 0.0068 0.0149 27.90 <.0001
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.0242 0.0057 -0.0355 -0.0130 17.90 <.0001
Ranked from 26 to 100 1 0.0109 0.0054 0.0004 0.0214 4.15 0.0417
Ranked from 11 to 25 1 0.0022 0.0075 -0.0125 0.0169 0.08 0.7711
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 -0.0028 0.0161 -0.0344 0.0289 0.03 0.8641
Behavioral 1 0.0072 0.0139 -0.0200 0.0344 0.27 0.6050
Biological 1 0.0231 0.0134 -0.0032 0.0494 2.97 0.0846
Chemistry 1 0.0334 0.0158 0.0024 0.0644 4.46 0.0348
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 1 0.2695 0.0015 0.2666 0.2725
Note: The scale parameter was estimated by maximum likelihood.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 112.66 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees 1 33.79 <.0001 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 1.71 0.1916 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 7.19 0.0073 LR
Kirschstein-NRSA postdoctorates versus other study groups 1 0.27 0.6015 LR
Table D.9. Regression Results Relating to the Years from Receipt of the Doctorate to First Competing R01 Awards
Model Information
Dataset WORK.PHD_GRANT_POSTDOC
Distribution Normal
Link function Identity
Dependent variable First R01 award received by Ph.D.’s
Observations
Number of observations read 39,055
Number of observations used 8,558
Missing values 30,497
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 8538 185959.9155 21.7803
Scaled deviance 8538 8558.0000 1.0023
Pearson Chi-Square 8538 185959.9155 21.7803
Scaled Pearson X2 8538 8558.0000 1.0023
Log likelihood -25316.8819
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 1.3175 0.9914 -0.6256 3.2607 1.77 0.1839
Ph.D. year 1 0.0653 0.0106 0.0444 0.0861 37.66 <.0001
Study group
Kirschstein-NRSA fellows 1 1.2185 0.3473 0.5379 1.8991 12.31 0.0005
Other non-Kirschstein-NRSA postdoctorates
1 1.3590 0.4136 0.5484 2.1697 10.80 0.0010
Kirschstein-NRSA trainees 1 2.0453 0.3423 1.3744 2.7162 35.70 <.0001
Kirschstein-NRSA trainee/fellows
1 1.6967 0.3744 0.9628 2.4306 20.53 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 1.3393 0.3448 0.6635 2.0151 15.09 0.0001
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
GenderFemale 1 0.7404 0.1132 0.5185 0.9623 42.76 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 0.4858 0.3195 -0.1404 1.1120 2.31 0.1284
Postdoctoral training 1 -0.7017 0.2622 -1.2157 -0.1877 7.16 0.0075
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 0.4953 0.1209 0.2585 0.7322 16.80 <.0001
9+ 1 0.4115 0.2612 -0.1005 0.9235 2.48 0.1152
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0101 0.0479 -0.0838 0.1041 0.04 0.8323
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 0.1482 0.1372 -0.1206 0.4171 1.17 0.2799
Ranked from 26 to 100 1 -0.0199 0.1243 -0.2634 0.2237 0.03 0.8731
Ranked from 11 to 25 1 -0.1651 0.1768 -0.5115 0.1813 0.87 0.3503
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 0.7791 0.4197 -0.0434 1.6016 3.45 0.0634
Behavioral 1 0.3306 0.3626 -0.3801 1.0412 0.83 0.3619
Biological 1 0.0302 0.3484 -0.6526 0.7131 0.01 0.9308
Chemistry 1 0.0585 0.3979 -0.7214 0.8385 0.02 0.8831
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 1 4.6615 0.0356 4.5922 4.7318
Note: The scale parameter was estimated by maximum likelihood.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 38.47 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees
1 24.04 <.0001 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 6.00 0.0143 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows
1 6.23 0.0125 LR
Kirschstein-NRSA postdoctorates versus other study groups 1 11.10 0.0009 LR
Table D.10. Regression Results Relating to the Years From Receipt of the Doctorate to First Competing R01-Equivalent Awards
Model Information
Dataset WORK.PHD_GRANT_POSTDOC1
Distribution Normal
Link function Identity
Dependent variable First R01-equivalent award received by Ph.D.’s
Observations
Number of observations read 15,830
Number of observations used 9,789
Missing values 6,041
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 9769 178477.0821 18.2697
Scaled deviance 9769 9789.0000 1.0020
Pearson Chi-Square 9769 178477.0821 18.2697
Scaled Pearson X2 9769 9789.0000 1.0020
Log likelihood -28099.7061
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 8.2147 0.8585 6.5322 9.8973 91.57 <.0001
Ph.D. year 1 -0.0324 0.0092 -0.0503 -0.0144 12.50 0.0004
Study group
Kirschstein-NRSA fellows 1 0.8995 0.3026 0.3063 1.4926 8.83 0.0030
Other non-Kirschstein-NRSA postdoctorates
1 1.1445 0.3652 0.4286 1.8603 9.82 0.0017
Kirschstein-NRSA trainees 1 1.4303 0.2980 0.8462 2.0145 23.03 <.0001
Kirschstein-NRSA trainee/fellows
1 1.0841 0.3252 0.4468 1.7214 11.12 0.0009
Unsuccessful Kirschstein-NRSA fellowship applicants
1 1.3654 0.3006 0.7763 1.9545 20.64 <.0001
Other research doctorates 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 0.6436 0.0961 0.4553 0.8319 44.88 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Expressed postdoctoral plans
Employment 1 0.6330 0.2763 0.0915 1.1744 5.25 0.0220
Postdoctoral training 1 -0.6837 0.2286 -1.1318 -0.2357 8.95 0.0028
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 0.3413 0.1044 0.1366 0.5460 10.68 0.0011
9+ 1 0.4523 0.2182 0.0245 0.8800 4.29 0.0382
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0633 0.0403 -0.0156 0.1423 2.47 0.1160
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 0.1129 0.1166 -0.1156 0.3415 0.94 0.3329
Ranked from 26 to 100 1 -0.1401 0.1070 -0.3498 0.0697 1.71 0.1906
Ranked from 11 to 25 1 -0.0904 0.1514 -0.3870 0.2063 0.36 0.5506
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Area of predoctoral study
Other field of science 1 0.4590 0.3557 -0.2382 1.1562 1.67 0.1969
Behavioral 1 0.6453 0.3093 0.0391 1.2516 4.35 0.0370
Biological 1 0.1670 0.2963 -0.4138 0.7477 0.32 0.5730
Chemistry 1 0.4103 0.3396 -0.2552 1.0758 1.46 0.2269
Health science 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 1 4.2699 0.0305 4.2105 4.3302
Note: The scale parameter was estimated by maximum likelihood.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 21.97 <.0001 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees 1 0.28 0.5957 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 1.33 0.2496 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA fellows 1 1.27 0.2589 LR
All Kirschstein-NRSA postdoctorates versus other study groups 1 6.04 0.0140 LR
Table D.11. Regression Results Reflecting the Likelihood of Kirschstein-NRSA Postdoctorates Submitting a Competing R01 Application, by Months of Support
Model Information
Dataset WORK.ALL_DEGREE_GRANTS
Distribution Binomial
Link function Logit
Dependent variable Applied for a R01 award, all degrees
Observations
Number of observations read 30,860
Number of observations used 30,729
Number of events 13,581
Number of trials 30,729
Missing values 131
Response Profile
OrderedValue R01applied Total
Frequency
1 1 13,581
2 0 17,148
PROC GENMOD is modeling the probability that Applied for a R01 award, all degrees = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 31E3 40566.1392 1.3209
Scaled deviance 31E3 40566.1392 1.3209
Pearson Chi-Square 31E3 30731.8433 1.0006
Scaled Pearson X2 31E3 30731.8433 1.0006
Log likelihood -20283.0696
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 3.4694 0.2584 2.9629 3.9759 180.25 <.0001
Last year of Kirschstein-NRSA postdoctoral support
1 -0.0472 0.0028 -0.0526 -0.0417 284.45 <.0001
Study group
Kirschstein-NRSA fellows 1 -0.0006 0.0481 -0.0950 0.0937 0.00 0.9895
Kirschstein-NRSA trainees 1 -0.5707 0.0453 -0.6596 -0.4819 158.64 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 0.1533 0.0337 0.0873 0.2193 20.75 <.0001
M.D./Ph.D. 1 0.3732 0.0452 0.2847 0.4617 68.27 <.0001
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Length of Kirschstein-NRSA support
<10 months 1 -0.3870 0.0766 -0.5371 -0.2368 25.51 <.0001
10–12 months 1 -0.4122 0.0604 -0.5306 -0.2938 46.56 <.0001
13–24 months 1 -0.1669 0.0569 -0.2783 -0.0554 8.62 0.0033
24–36 months 1 0.0555 0.0556 -0.0534 0.1644 1.00 0.3178
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2936 0.0262 -0.3449 -0.2424 126.06 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.2594 0.0380 -0.3338 -0.1850 46.69 <.0001
Ranked from 26 to 100 1 -0.1296 0.0304 -0.1892 -0.0701 18.19 <.0001
Ranked from 11 to 25 1 -0.1356 0.0335 -0.2013 -0.0699 16.37 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
field2 Other field of science 1 0.8879 0.0934 0.7049 1.0709 90.43 <.0001
field2 behavioral 1 0.8698 0.1014 0.6710 1.0686 73.55 <.0001
field2 biological 1 1.0440 0.0919 0.8639 1.2242 129.04 <.0001
field2 chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 393.61 <.0001 LR
M.D.’s versus Ph.D.’s 1 20.74 <.0001 LR
Table D.12. Regression Results Reflecting the Likelihood of Kirschstein-NRSA Postdoctorates Submitting a Competing R01-Equivalent Application, by Months of Support
Model Information
Dataset WORK.ALL_DEGREE_GRANTS
Distribution Binomial
Link function Logit
Dependent variable Applied for a R01-equivalent award, all degrees
Observations
Number of observations read 30,860
Number of observations used 30,729
Number of events 15,044
Number of trials 30,729
Missing values 131
Response Profile
OrderedValue R01r23r29applied Total
Frequency
1 1 15044
2 0 15685
PROC GENMOD is modeling the probability that Applied for a R01-equivalent award, all degrees = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 31E3 40997.7815 1.3349
Scaled deviance 31E3 40997.7815 1.3349
Pearson Chi-Square 31E3 30737.6516 1.0008
Scaled Pearson X2 31E3 30737.6516 1.0008
Log likelihood -20498.8907
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 2.7565 0.2561 2.2546 3.2583 115.89 <.0001
Last year of Kirschstein-NRSA postdoctoral study
1 -0.0368 0.0028 -0.0422 -0.0313 176.33 <.0001
Study group
Kirschstein-NRSA fellows 1 -0.0194 0.0489 -0.1152 0.0765 0.16 0.6922
Kirschstein-NRSA trainees 1 -0.5778 0.0459 -0.6678 -0.4877 158.16 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 0.1952 0.0333 0.1298 0.2605 34.29 <.0001
M.D./Ph.D. 1 0.3676 0.0453 0.2787 0.4565 65.72 <.0001
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Length of Kirschstein-NRSA support
<10 months 1 -0.4993 0.0763 -0.6488 -0.3497 42.81 <.0001
10–12 months 1 -0.4995 0.0604 -0.6179 -0.3811 68.37 <.0001
13–24 months 1 -0.2078 0.0571 -0.3197 -0.0960 13.26 0.0003
24–36 months 1 0.0338 0.0560 -0.0759 0.1435 0.37 0.5453
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2735 0.0258 -0.3241 -0.2229 112.19 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.2667 0.0377 -0.3405 -0.1928 50.12 <.0001
Ranked from 26 to 100 1 -0.1298 0.0302 -0.1890 -0.0705 18.42 <.0001
Ranked from 11 to 25 1 -0.1324 0.0333 -0.1977 -0.0672 15.83 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 0.9628 0.0919 0.7828 1.1428 109.86 <.0001
Behavioral 1 0.9391 0.0997 0.7436 1.1346 88.65 <.0001
Biological 1 1.1229 0.0904 0.9456 1.3002 154.13 <.0001
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 378.87 <.0001 LR
M.D.’s versus Ph.D.’s 1 34.34 <.0001 LR
Table D.13. Regression Results Reflecting the Likelihood of Kirschstein-NRSA Postdoctorates Being Awarded a Competing R01 Application, by Months of Support
Model Information
Dataset WORK.POSTDOC1
Distribution Binomial
Link function Logit
Dependent variable Awarded a R01 award
Observations
Number of observations read 13,626
Number of observations used 13,581
Number of events 8,203
Number of trials 13,581
Missing values 45
Response Profile
OrderedValue R01awarded Total
Frequency
1 1 8,203
2 0 5,378
PROC GENMOD is modeling the probability that Awarded a R01 award = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 14E3 17876.0051 1.3179
Scaled deviance 14E3 17876.0051 1.3179
Pearson Chi-Square 14E3 13576.2007 1.0009
Scaled Pearson X2 14E3 13576.2007 1.0009
Log likelihood -8938.0026
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 2.6040 0.4054 1.8095 3.3985 41.27 <.0001
Last year of Kirschstein-NRSA postdoctoral support
1 -0.0207 0.0043 -0.0292 -0.0123 23.16 <.0001
Study group
Kirschstein-NRSA fellows 1 0.1322 0.0671 0.0007 0.2637 3.88 0.0488
Kirschstein-NRSA trainees 1 -0.3829 0.0636 -0.5075 -0.2582 36.24 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 -0.0548 0.0511 -0.1549 0.0454 1.15 0.2836
M.D./Ph.D. 1 0.1109 0.0654 -0.0172 0.2391 2.88 0.0898
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Length of Kirschstein-NRSA support
<10 months 1 0.1069 0.1166 -0.1216 0.3354 0.84 0.3593
10–12 months 1 -0.0000 0.0885 -0.1735 0.1735 0.00 1.0000
13–24 months 1 0.0393 0.0821 -0.1216 0.2002 0.23 0.6323
24–36 months 1 0.0465 0.0794 -0.1091 0.2021 0.34 0.5581
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2366 0.0403 -0.3155 -0.1576 34.50 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.4299 0.0577 -0.5429 -0.3169 55.59 <.0001
Ranked from 26 to 100 1 -0.2670 0.0458 -0.3568 -0.1772 33.94 <.0001
Ranked from 11 to 25 1 -0.2369 0.0507 -0.3363 -0.1376 21.84 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 -0.0427 0.1650 -0.3660 0.2807 0.07 0.7958
Behavioral 1 -0.3623 0.1762 -0.7077 -0.0169 4.23 0.0398
Biological 1 0.0442 0.1629 -0.2750 0.3634 0.07 0.7862
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 146.76 <.0001 LR
M.D.’s versus Ph.D.’s 1 1.15 0.2838 LR
Table D.14. Regression Results Reflecting the Likelihood of Kirschstein-NRSA Postdoctorates Being Awarded a Competing R01-Equivalent Application, by Months of Support
Model Information
Dataset WORK.POSTDOC1
Distribution Binomial
Link function Logit
Dependent variable Awarded a R01-equivalent award
Observations
Number of observations read 15,092
Number of observations used 15,044
Number of events 9,445
Number of trials 15,044
Missing values 48
Response Profile
OrderedValue R01r23r29awarded Total
Frequency
1 1 9,445
2 0 5,599
PROC GENMOD is modeling the probability that Awarded a R01-equivalent award = “1.'”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 15E3 19458.0726 1.2949
Scaled deviance 15E3 19458.0726 1.2949
Pearson Chi-Square 15E3 15040.9740 1.0009
Scaled Pearson X2 15E3 15040.9740 1.0009
Log likelihood -9729.0363
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 2.3216 0.3921 1.5532 3.0901 35.06 <.0001
Last year of Kirschstein-NRSA postdoctoral support
1 -0.0157 0.0041 -0.0238 -0.0076 14.37 0.0002
Study group
Kirschstein-NRSA fellows 1 0.1068 0.0655 -0.0216 0.2352 2.66 0.1031
Kirschstein-NRSA trainees 1 -0.4130 0.0617 -0.5339 -0.2921 44.83 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 -0.0898 0.0488 -0.1853 0.0058 3.39 0.0656
M.D./Ph/D. 1 0.1322 0.0636 0.0075 0.2568 4.32 0.0377
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Length of Kirschstein-NRSA support
<10 months 1 0.0526 0.1128 -0.1685 0.2736 0.22 0.6412
10–12 months 1 -0.0452 0.0850 -0.2117 0.1214 0.28 0.5949
13–24 months 1 0.0329 0.0788 -0.1216 0.1875 0.17 0.6761
24–36 months 1 0.0521 0.0763 -0.0975 0.2017 0.47 0.4948
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2265 0.0384 -0.3018 -0.1512 34.76 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.4031 0.0554 -0.5117 -0.2944 52.84 <.0001
Ranked from 26 to 100 1 -0.2418 0.0441 -0.3282 -0.1554 30.08 <.0001
Ranked from 11 to 25 1 -0.2054 0.0487 -0.3009 -0.1099 17.77 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 -0.0659 0.1631 -0.3856 0.2538 0.16 0.6864
Behavioral 1 -0.4049 0.1733 -0.7446 -0.0653 5.46 0.0195
Biological 1 0.0539 0.1612 -0.2620 0.3698 0.11 0.7380
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 159.82 <.0001 LR
M.D.’s versus Ph.D.’s 1 3.38 0.0659 LR
Table D.15. Regression Results of the Number of R01 Awards, Given the Submission of a Competing R01 Application
Model Information
Dataset WORK.POSTDOC1
Distribution Poisson
Link function Log
Dependent variable Number of R01 awards
Observations
Number of observations read 13,626
Number of observations used 13,581
Missing values 45
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 14E3 34853.3154 2.5695
Scaled deviance 14E3 34853.3154 2.5695
Pearson Chi-Square 14E3 35894.2465 2.6463
Scaled Pearson X2 14E3 35894.2465 2.6463
Log likelihood -8387.7276
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 5.7085 0.1453 5.4238 5.9932 1544.18 <.0001
Last year of Kirschstein-NRSA postdoctoral support
1 -0.0565 0.0016 -0.0597 -0.0534 1234.49 <.0001
Study group
Kirschstein-NRSA fellows 1 0.0592 0.0224 0.0153 0.1031 6.98 0.0082
Kirschstein-NRSA trainees 1 -0.2563 0.0221 -0.2996 -0.2129 134.22 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 -0.0252 0.0192 -0.0628 0.0123 1.74 0.1878
M.D./Ph.D. 1 0.1507 0.0226 0.1064 0.1950 44.50 <.0001
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Length of Kirschstein-NRSA support
<10 months 1 0.1136 0.0402 0.0347 0.1925 7.97 0.0048
10–12 months 1 -0.0439 0.0324 -0.1074 0.0197 1.83 0.1760
13–24 months 1 -0.0203 0.0296 -0.0784 0.0378 0.47 0.4934
24–36 months 1 0.0163 0.0285 -0.0396 0.0722 0.33 0.5680
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.1905 0.0155 -0.2208 -0.1602 151.57 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.2568 0.0207 -0.2973 -0.2163 154.39 <.0001
Ranked from 26 to 100 1 -0.1478 0.0159 -0.1790 -0.1166 86.33 <.0001
Ranked from 11 to 25 1 -0.1518 0.0180 -0.1871 -0.1165 71.13 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 -0.1560 0.0490 -0.2521 -0.0599 10.12 0.0015
Behavioral 1 -0.4432 0.0567 -0.5544 -0.3320 61.03 <.0001
Biological 1 -0.0548 0.0477 -0.1483 0.0387 1.32 0.2504
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 453.83 <.0001 LR
M.D.’s versus Ph.D.’s 1 1.74 0.1872 LR
Table D.16. Regression Results of the Number of R01-Equivalent Awards, Given the Submission of a Competing R01-Equivalent Application
Model Information
Dataset WORK.POSTDOC1
Distribution Poisson
Link function Log
Dependent variable Number of R01-equivalent awards
Observations
Number of observations read 15,092
Number of observations used 15,044
Missing values 48
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 15E3 37554.3099 2.4991
Scaled deviance 15E3 37554.3099 2.4991
Pearson Chi-Square 15E3 38599.5529 2.5687
Scaled Pearson X2 15E3 38599.5529 2.5687
Log likelihood -9175.1130
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 5.2314 0.1364 4.9642 5.4987 1471.65 <.0001
Last year of Kirschstein-NRSA postdoctoral support
1 -0.0503 0.0015 -0.0532 -0.0473 1115.84 <.0001
Study group
Kirschstein-NRSA fellows 1 0.0504 0.0211 0.0090 0.0917 5.69 0.0170
Kirschstein-NRSA trainees 1 -0.2670 0.0207 -0.3076 -0.2263 165.73 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 -0.0413 0.0180 -0.0765 -0.0060 5.25 0.0219
M.D./Ph.D. 1 0.1445 0.0213 0.1028 0.1863 46.04 <.0001
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Length of Kirschstein-NRSA support
<10 months 1 0.1077 0.0382 0.0329 0.1825 7.97 0.0048
10–12 months 1 -0.0405 0.0303 -0.0999 0.0189 1.78 0.1818
13–24 months 1 -0.0176 0.0276 -0.0718 0.0365 0.41 0.5230
24–36 months 1 0.0199 0.0266 -0.0321 0.0720 0.56 0.4532
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.1852 0.0144 -0.2134 -0.1569 165.09 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.2520 0.0195 -0.2903 -0.2137 166.18 <.0001
Ranked from 26 to 100 1 -0.1382 0.0150 -0.1676 -0.1088 84.83 <.0001
Ranked from 11 to 25 1 -0.1425 0.0169 -0.1757 -0.1093 70.80 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 -0.1648 0.0470 -0.2570 -0.0727 12.29 0.0005
Behavioral 1 -0.4464 0.0541 -0.5524 -0.3404 68.12 <.0001
Biological 1 -0.0676 0.0458 -0.1574 0.0222 2.18 0.1402
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 515.30 <.0001 LR
M.D.’s versus Ph.D.’s 1 5.27 0.0216 LR
Table D.17. Results From a Regression Analysis of the Average Award Rate (Proportion of Grants Awarded to Grants Competed for Kirschstein-NRSA Postdoctoral Recipients Only) for Competing R01 Grants
Model Information
Dataset WORK.POSTDOC1
Distribution Normal
Link function Identity
Dependent variable R01 award rate
Observations
Number of observations read 13,626
Number of observations used 13,581
Missing values 45
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 14E3 986.6419 0.0727
Scaled deviance 14E3 13581.0000 1.0013
Pearson Chi-Square 14E3 986.6419 0.0727
Scaled Pearson X2 14E3 13581.0000 1.0013
Log likelihood -1465.0991
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 0.3141 0.0525 0.2113 0.4170 35.81 <.0001
Last year of Kirschstein-NRSA postdoctoral support
1 -0.0003 0.0006 -0.0014 0.0008 0.29 0.5931
Study group
Kirschstein-NRSA fellows 1 0.0220 0.0086 0.0052 0.0388 6.56 0.0104
Kirschstein-NRSA fellows 1 -0.0420 0.0082 -0.0581 -0.0259 26.22 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 -0.0072 0.0067 -0.0204 0.0059 1.17 0.2804
M.D./Ph.D. 1 0.0143 0.0085 -0.0022 0.0309 2.87 0.0902
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Length of Kirschstein-NRSA support
<10 months 1 0.0253 0.0152 -0.0044 0.0550 2.79 0.0947
10–12 months 1 0.0083 0.0116 -0.0144 0.0309 0.51 0.4751
13–24 months 1 0.0026 0.0107 -0.0184 0.0235 0.06 0.8096
24–36 months 1 -0.0028 0.0103 -0.0231 0.0174 0.08 0.7827
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.0225 0.0053 -0.0329 -0.0121 18.05 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.0503 0.0075 -0.0649 -0.0356 45.13 <.0001
Ranked from 26 to 100 1 -0.0423 0.0059 -0.0539 -0.0308 51.34 <.0001
Ranked from 11 to 25 1 -0.0354 0.0066 -0.0483 -0.0226 29.12 <.0001
Top 10 institution 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 0.0071 0.0208 -0.0336 0.0478 0.12 0.7325
Behavioral 1 -0.0305 0.0224 -0.0743 0.0134 1.86 0.1731
Biological 1 0.0111 0.0205 -0.0290 0.0512 0.30 0.5870
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 1 0.2695 0.0016 0.2663 0.2728
Note: The scale parameter was estimated by maximum likelihood.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 134.68 <.0001 LR
M.D.’s versus Ph.D.’s 1 1.16 0.2804 LR
Table D.18. Results From a Regression Analysis of the Average Award Rate (Proportion of Grants Awarded to Grants Competed for Kirschstein-NRSA Postdoctoral Recipients Only) for Competing R01-Equivalent Grants
Model Information
Dataset WORK.POSTDOC1
Distribution Normal
Link function Identity
Dependent variable R01-equivalent award rate
Observations
Number of observations read 15,092
Number of observations used 15,044
Missing values 48
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 15E3 1097.1527 0.0730
Scaled deviance 15E3 15044.0000 1.0011
Pearson Chi-Square 15E3 1097.1527 0.0730
Scaled Pearson X2 15E3 15044.0000 1.0011
Log likelihood -1651.9533
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 0.3745 0.0502 0.2760 0.4729 55.56 <.0001
Last year of Kirschstein-NRSA postdoctoral support
1 -0.0008 0.0005 -0.0019 0.0002 2.32 0.1279
Study group
Kirschstein-NRSA fellows 1 0.0163 0.0082 0.0002 0.0324 3.96 0.0467
Kirschstein-NRSA trainees 1 -0.0441 0.0078 -0.0594 -0.0288 31.83 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 -0.0079 0.0064 -0.0204 0.0046 1.54 0.2144
M.D./Ph.D. 1 0.0145 0.0081 -0.0014 0.0303 3.20 0.0735
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Length of Kirschstein-NRSA support
<10 months 1 0.0150 0.0145 -0.0135 0.0435 1.06 0.3024
10–12 months 1 0.0002 0.0110 -0.0213 0.0217 0.00 0.9837
13–24 months 1 0.0002 0.0101 -0.0197 0.0200 0.00 0.9861
24–36 months 1 -0.0054 0.0098 -0.0246 0.0137 0.31 0.5779
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.0218 0.0050 -0.0316 -0.0119 18.90 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.0496 0.0071 -0.0636 -0.0356 48.42 <.0001
Ranked from 26 to 100 1 -0.0398 0.0056 -0.0508 -0.0288 50.04 <.0001
Ranked from 11 to 25 1 -0.0303 0.0062 -0.0425 -0.0180 23.49 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 0.0027 0.0202 -0.0369 0.0423 0.02 0.8932
Behavioral 1 -0.0358 0.0217 -0.0783 0.0067 2.72 0.0991
Biological 1 0.0091 0.0199 -0.0300 0.0481 0.21 0.6492
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 1 0.2701 0.0016 0.2670 0.2731
Note: The scale parameter was estimated by maximum likelihood.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 131.66 <.0001 LR
M.D.’s versus Ph.D.’s 1 1.54 0.2144 LR
Table D.19. Years to First Award of an R01 Competing Grant
Model Information
Dataset WORK.POSTDOC1
Distribution Normal
Link function Identity
Dependent variable Years to first R01 award
Observations
Number of observations read 13,626
Number of observations used 8,181
Missing values 5,445
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 8164 145096.1820 17.7727
Scaled deviance 8164 8181.0000 1.0021
Pearson Chi-Square 8164 145096.1820 17.7727
Scaled Pearson X2 8164 8181.0000 1.0021
Log likelihood -23370.9062
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 -4.7336 1.0579 -6.8072 -2.6601 20.02 <.0001
Last year of Kirschstein-NRSA postdoctoral support
1 0.1186 0.0115 0.0961 0.1410 106.75 <.0001
Study group
Kirschstein-NRSA fellows 1 0.0709 0.1661 -0.2547 0.3964 0.18 0.6697
Kirschstein-NRSA trainees 1 0.9765 0.1630 0.6570 1.2960 35.88 <.0001
Kirschstein-NRSA fellow/trainees
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 0.8559 0.1388 0.5838 1.1279 38.03 <.0001
M.D./Ph.D. 1 -0.0740 0.1683 -0.4038 0.2558 0.19 0.6602
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Length of Kirschstein-NRSA support
<10 months 1 0.2901 0.3047 -0.3071 0.8873 0.91 0.3410
10–12 months 1 0.7890 0.2346 0.3292 1.2487 11.31 0.0008
13–24 months 1 0.2959 0.2147 -0.1250 0.7167 1.90 0.1682
24–36 months 1 0.0275 0.2073 -0.3788 0.4338 0.02 0.8945
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 0.6078 0.1094 0.3933 0.8223 30.84 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 0.1738 0.1530 -0.1260 0.4736 1.29 0.2558
Ranked from 26 to 100 1 0.0819 0.1174 -0.1482 0.3120 0.49 0.4855
Ranked from 11 to 25 1 0.1999 0.1315 -0.0579 0.4577 2.31 0.1285
Top 10 institution 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 0.2215 0.4012 -0.5649 1.0078 0.30 0.5810
Behavioral 1 0.4723 0.4446 -0.3991 1.3437 1.13 0.2881
Biological 1 -0.3576 0.3937 -1.1293 0.4141 0.82 0.3638
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 1 4.2114 0.0329 4.1473 4.2764
Note: The scale parameter was estimated by maximum likelihood.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 66.43 <.0001 LR
M.D.’s versus Ph.D.’s 1 37.94 <.0001 LR
Table D.20. Years to First Award of an R01-Equivalent Competing Grant
Model Information
Dataset WORK.POSTDOC1
Distribution Normal
Link function Identity
Dependent variable Years to first R01-equivalent award
Observations
Number of observations read 15,092
Number of observations used 9,588
Missing values 5,504
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 9571 161177.5191 16.8402
Scaled deviance 9571 9588.0000 1.0018
Pearson Chi-Square 9571 161177.5191 16.8402
Scaled Pearson X2 9571 9588.0000 1.0018
Log likelihood -27133.4221
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 0.2798 0.9554 -1.5926 2.1523 0.09 0.7696
Last year of Kirschstein-NRSA postdoctoral support
1 0.0361 0.0103 0.0158 0.0563 12.20 0.0005
Study group
Kirschstein-NRSA fellows 1 0.1885 0.1495 -0.1046 0.4815 1.59 0.2075
Kirschstein-NRSA trainees 1 0.8861 0.1454 0.6011 1.1710 37.14 <.0001
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
Degree
M.D. 1 1.0962 0.1247 0.8519 1.3405 77.32 <.0001
M.D./Ph.D. 1 0.0721 0.1511 -0.2240 0.3682 0.23 0.6333
Ph.D. 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Length of Kirschstein-NRSA support
<10 months 1 0.9101 0.2745 0.3722 1.4481 11.00 0.0009
10–12 months 1 1.1149 0.2081 0.7071 1.5226 28.71 <.0001
13–24 months 1 0.6485 0.1897 0.2767 1.0202 11.69 0.0006
24–36 months 1 0.2850 0.1831 -0.0738 0.6439 2.42 0.1195
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 0.4500 0.0972 0.2595 0.6405 21.44 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 0.0442 0.1371 -0.2246 0.3130 0.10 0.7470
Ranked from 26 to 100 1 -0.1075 0.1060 -0.3154 0.1003 1.03 0.3104
Ranked from 11 to 25 1 0.0163 0.1183 -0.2156 0.2481 0.02 0.8907
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Field of postdoctoral study
Other field of science 1 0.1624 0.3713 -0.5653 0.8902 0.19 0.6617
Behavioral 1 0.2386 0.4077 -0.5605 1.0377 0.34 0.5585
Biological 1 -0.1425 0.3649 -0.8578 0.5728 0.15 0.6962
Chemistry 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 1 4.1000 0.0296 4.0424 4.1585
Note: The scale parameter was estimated by maximum likelihood.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 48.94 <.0001 LR
M.D.’s versus Ph.D.’s 1 77.01 <.0001 LR
Table D.21. Number of Publications After Receipt of Doctorate
Model Information
Dataset WORK.ISI_POSTDOC
Distribution Poisson
Link function Log
Dependent variable Number of publications completed for all Ph.D.’s
Observations
Number of observations read 2,531
Number of observations used 2,517
Missing values 14
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 2500 20666.8826 8.2668
Scaled deviance 2500 20666.8826 8.2668
Pearson Chi-Square 2500 25892.0444 10.3568
Scaled Pearson X2 2500 25892.0444 10.3568
Log likelihood 22064.2792
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 1.3056 0.0730 1.1626 1.4486 320.10 <.0001
Cohort81/82 1 0.5972 0.0165 0.5648 0.6295 1307.26 <.0001
87/88 0 0.0000 0.0000 0.0000 0.0000 . .
Study group
Kirschstein-NRSA fellows
1 0.4949 0.0314 0.4334 0.5563 249.01 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.2898 0.0276 0.2358 0.3438 110.55 <.0001
Kirschstein-NRSA trainees
1 0.4750 0.0268 0.4224 0.5275 313.94 <.0001
Kirschstein-NRSA trainee/fellows
1 0.4803 0.0401 0.4017 0.5589 143.34 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 0.3109 0.0295 0.2531 0.3688 110.96 <.0001
Other research doctorates
0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.4994 0.0160 -0.5307 -0.4682 979.51 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 -0.0132 0.0693 -0.1490 0.1226 0.04 0.8488
Postdoctoral training 1 0.3317 0.0673 0.1999 0.4636 24.33 <.0001
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.0700 0.0191 -0.1074 -0.0326 13.45 0.0002
9+ 1 -0.2854 0.0275 -0.3393 -0.2315 107.73 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support
1 -0.0101 0.0065 -0.0228 0.0025 2.46 0.1167
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 0.0273 0.0193 -0.0106 0.0652 1.99 0.1579
Ranked from 26 to 100 1 0.0001 0.0184 -0.0360 0.0361 0.00 0.9971
Ranked from 11 to 25 1 0.0423 0.0266 -0.0099 0.0944 2.52 0.1122
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Publications count during predoctoral work
1 0.1882 0.0041 0.1801 0.1962 2097.44 <.0001
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 0.15 0.7019 LR
Unsuccessful fellowship applicants versus trainees 1 133.90 <.0001 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 24.47 <.0001 LR
Table D.22. Number of Publications After Receipt of Doctorate
Model Information
Dataset WORK.ISI_POSTDOC2
Distribution Poisson
Link function Log
Dependent variable Number of publications completed for postdoctorate population
Observations
Number of observations read 883
Number of observations used 883
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 861 7831.9536 9.0963
Scaled deviance 861 7831.9536 9.0963
Pearson Chi-Square 861 8736.0643 10.1464
Scaled Pearson X2 861 8736.0643 10.1464
Log likelihood 12695.4418
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 1.7547 0.1299 1.5002 2.0093 182.49 <.0001
Cohort81/82 1 0.7279 0.0266 0.6758 0.7800 750.16 <.0001
87/88 0 0.0000 0.0000 0.0000 0.0000 . .
Study group
Kirschstein-NRSA fellows 1 0.0779 0.0409 -0.0022 0.1580 3.63 0.0567
Kirschstein-NRSA trainees 1 -0.0066 0.0383 -0.0817 0.0685 0.03 0.8636
Kirschstein-NRSA trainee/fellows
0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.4981 0.0231 -0.5434 -0.4528 465.41 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 0.1522 0.1213 -0.0855 0.3899 1.57 0.2095
Postdoctoral training 1 0.3385 0.1134 0.1163 0.5606 8.92 0.0028
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Elapsed years enrolled in doctoral program
5–8 1 -0.1071 0.0274 -0.1609 -0.0534 15.27 <.0001
9+ 1 -0.2974 0.0436 -0.3828 -0.2120 46.56 <.0001
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support 1 0.0042 0.0070 -0.0094 0.0179 0.37 0.5416
NIH extramural funding (ranking of postdoctoral institution)
Ranked 101 or more 1 -0.2550 0.0345 -0.3227 -0.1873 54.47 <.0001
Ranked from 26 to 100 1 0.0558 0.0273 0.0023 0.1094 4.18 0.0408
Ranked from 11 to 25 1 -0.1749 0.0317 -0.2371 -0.1127 30.34 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 0.0590 0.0288 0.0025 0.1155 4.20 0.0405
Ranked from 26 to 100 1 -0.0326 0.0270 -0.0855 0.0203 1.46 0.2267
Ranked from 11 to 25 1 0.2350 0.0396 0.1574 0.3127 35.18 <.0001
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Postdoctoral field of study
Other field of science 1 -0.1551 0.0250 -0.2042 -0.1060 38.37 <.0001
biological 0 0.0000 0.0000 0.0000 0.0000 . .
Publications count during predoctoral work 1 0.1930 0.0082 0.1769 0.2092 549.52 <.0001
Length of Kirschstein-NRSA support
<10 months 1 -0.3376 0.0779 -0.4902 -0.1849 18.78 <.0001
10–12 months 1 0.0468 0.0518 -0.0546 0.1483 0.82 0.3658
13–24 months 1 -0.0664 0.0474 -0.1593 0.0265 1.96 0.1614
24–36 months 1 0.0979 0.0454 0.0089 0.1869 4.65 0.0311
37+ months 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 10.89 0.0010 LR
Table D.23. Regression Results of the Proportion of Individuals Employed at a University or Affiliated Research Facility 8 Years After Receipt of the Doctorate
Model Information
Dataset WORK.EMPLOY_ORG
Distribution Binomial
Link function Logit
Dependent variable Employment at university 8 years after receipt of the doctorate
Observations
Number of observations read 10,659
Number of observations used 7,660
Number of events 3,704
Number of trials 7,660
Missing values 2,999
Response Profile
OrderedValue empl_type2 Total
Frequency
1 1 3704
2 0 3956
PROC GENMOD is modeling the probability that Employment at university 8 years after receipt of the doctorate = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 7644 10386.2311 1.3587
Scaled deviance 7644 10386.2311 1.3587
Pearson Chi-Square 7644 7661.7306 1.0023
Scaled Pearson X2 7644 7661.7306 1.0023
Log likelihood -5193.1155
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 1.6115 0.3250 0.9746 2.2485 24.59 <.0001
Ph.D. year 1 -0.0256 0.0037 -0.0329 -0.0184 47.95 <.0001
Study group
Kirschstein-NRSA fellows
1 0.7986 0.1223 0.5588 1.0383 42.62 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.1841 0.0733 0.0404 0.3278 6.31 0.0120
Kirschstein-NRSA trainees
1 0.5402 0.1066 0.3312 0.7492 25.67 <.0001
Kirschstein-NRSA trainee/fellows
1 0.7418 0.1890 0.3713 1.1123 15.40 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 0.5199 0.1144 0.2956 0.7441 20.65 <.0001
Other research doctorates
0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 0.2418 0.0492 0.1453 0.3383 24.11 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 0.1211 0.1515 -0.1759 0.4181 0.64 0.4243
Postdoctoral training 1 0.1436 0.1542 -0.1585 0.4457 0.87 0.3516
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.0115 0.0578 -0.1248 0.1019 0.04 0.8427
9+ 1 -0.0789 0.0786 -0.2330 0.0752 1.01 0.3157
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support
1 0.0763 0.0406 -0.0032 0.1558 3.53 0.0601
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.1376 0.0600 -0.2551 -0.0200 5.26 0.0219
Ranked from 26 to 100
1 0.0860 0.0592 -0.0301 0.2020 2.11 0.1465
Ranked from 11 to 25 1 0.2150 0.0817 0.0548 0.3751 6.92 0.0085
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 4.26 0.0390 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees
1 0.03 0.8741 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 13.03 0.0003 LR
Table D.24. Regression Results of the Proportion of Individuals Employed at a University or Affiliated Research Facility 12 Years After Receipt of the Doctorate
Model Information
Dataset WORK.EMPLOY_ORG
Distribution Binomial
Link function Logit
Dependent variable Employment at university 12 years after receipt of the doctorate
Observations
Number of observations read 10,659
Number of observations used 7,290
Number of events 3,390
Number of trials 7,290
Missing values 3,369
Response Profile
OrderedValue empl_type3 Total
Frequency
1 1 3390
2 0 3900
PROC GENMOD is modeling the probability that Employment at university 12 years after receipt of the doctorate = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 7274 9880.2782 1.3583
Scaled deviance 7274 9880.2782 1.3583
Pearson Chi-Square 7274 7290.1413 1.0022
Scaled Pearson X2 7274 7290.1413 1.0022
Log likelihood -4940.1391
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 1.7107 0.3466 1.0315 2.3900 24.37 <.0001
Ph.D. year 1 -0.0281 0.0041 -0.0362 -0.0200 46.44 <.0001
Study group
Kirschstein-NRSA fellows
1 0.6294 0.1225 0.3894 0.8695 26.40 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 0.2750 0.0744 0.1293 0.4208 13.68 0.0002
Kirschstein-NRSA trainees
1 0.5847 0.1072 0.3746 0.7948 29.76 <.0001
Kirschstein-NRSA trainee/fellows
1 0.7457 0.1865 0.3801 1.1113 15.98 <.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 0.6002 0.1176 0.3697 0.8307 26.04 <.0001
Other research doctorates
0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 0.1817 0.0509 0.0820 0.2814 12.76 0.0004
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 0.1756 0.1567 -0.1316 0.4828 1.25 0.2627
Postdoctoral training 1 0.1681 0.1590 -0.1436 0.4798 1.12 0.2906
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.0668 0.0585 -0.1815 0.0479 1.30 0.2536
9+ 1 -0.0877 0.0826 -0.2495 0.0742 1.13 0.2883
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support
1 0.1096 0.0411 0.0290 0.1901 7.10 0.0077
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.1520 0.0622 -0.2740 -0.0300 5.96 0.0146
Ranked from 26 to 100 1 0.0442 0.0597 -0.0728 0.1613 0.55 0.4587
Ranked from 11 to 25 1 0.0757 0.0848 -0.0906 0.2419 0.80 0.3725
Top 10 institution 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 0.13 0.7189 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees
1 0.01 0.9048 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 9.70 0.0018 LR
Table D.25. Regression Results of the Proportion of Individuals Employed as an Associate or Full Professor 8 Years After Receipt of the Doctorate
Model Information
Dataset WORK.RANK
Distribution Binomial
Link function Logit
Dependent variable Employment as associate or full professor 8 years after doctorate
Observations
Number of observations read 10,659
Number of observations used 2,929
Number of events 1,383
Number of trials 2,929
Missing values 7,730
Response Profile
OrderedValue fac_rank2 Total
Frequency
1 1 1383
2 0 1546
PROC GENMOD is modeling the probability that Employment as associate or full professor 8 years after doctorate = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 2913 3218.8721 1.1050
Scaled deviance 2913 3218.8721 1.1050
Pearson Chi-Square 2913 2935.6323 1.0078
Scaled Pearson X2 2913 2935.6323 1.0078
Log likelihood -1609.4361
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 4.5543 0.6057 3.3672 5.7414 56.54 <.0001
Ph.D. year 1 -0.0540 0.0072 -0.0680 -0.0399 56.61 <.0001
Study group
Kirschstein-NRSA fellows
1 -1.2152 0.2169 -1.6403 -0.7901 31.39 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 -0.5951 0.1289 -0.8477 -0.3425 21.32 <.0001
Kirschstein-NRSA trainees
1 -1.0655 0.1933 -1.4444 -0.6866 30.38 <.0001
Kirschstein-NRSA fellow/trainees
1 -1.5240 0.4017 -2.3113 -0.7367 14.40 0.0001
Unsuccessful Kirschstein-NRSA fellowship applicants
1 -1.1315 0.2179 -1.5586 -0.7044 26.96 <.0001
Other research doctorates
0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.4090 0.0935 -0.5923 -0.2258 19.15 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 0.8121 0.2666 0.2896 1.3347 9.28 0.0023
Postdoctoral 1 -0.4286 0.2731 -0.9639 0.1067 2.46 0.1165
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.1540 0.1054 -0.3606 0.0526 2.13 0.1440
9+ 1 -0.0327 0.1498 -0.3263 0.2609 0.05 0.8271
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support
1 -0.1557 0.0925 -0.3369 0.0256 2.83 0.0923
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 0.1488 0.1145 -0.0755 0.3732 1.69 0.1935
Ranked from 26 to 100 1 -0.0066 0.1080 -0.2183 0.2051 0.00 0.9513
Ranked from 11 to 25 1 0.2051 0.1495 -0.0880 0.4982 1.88 0.1702
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 0.40 0.5263 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees
1 0.07 0.7909 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 6.57 0.0104 LR
Table D.26. Regression Results of the Proportion of Individuals Employed as an Associate or Full Professor 12 Years After Receipt of the Doctorate
Model Information
Dataset WORK.RANK
Distribution Binomial
Link function Logit
Dependent variable Employment as associate or full professor 12 years after receipt of the doctorate
Observations
Number of observations read 10,659
Number of observations used 2,598
Number of events 2,045
Number of trials 2,598
Missing values 8,061
Response Profile
OrderedValue fac_rank3 Total
Frequency
1 1 2,045
2 0 553
PROC GENMOD is modeling the probability that Employment as associate or full professor 12 years after receipt of the doctorate = “1.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 2582 2331.4572 0.9030
Scaled deviance 2582 2331.4572 0.9030
Pearson Chi-Square 2582 2572.0329 0.9961
Scaled Pearson X2 2582 2572.0329 0.9961
Log likelihood -1165.7286
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 5.8502 0.8797 4.1260 7.5745 44.22 <.0001
Ph.D. year 1 -0.0575 0.0111 -0.0792 -0.0358 27.01 <.0001
Study group
Kirschstein-NRSA fellows
1 -0.4733 0.2243 -0.9130 -0.0336 4.45 0.0349
Other non-Kirschstein-NRSA postdoctorates
1 -0.3829 0.1743 -0.7244 -0.0413 4.83 0.0280
Kirschstein-NRSA trainees
1 -0.9571 0.2054 -1.3597 -0.5546 21.72 <.0001
Kirschstein-NRSA fellow/trainees
1 -0.7034 0.3059 -1.3030 -0.1039 5.29 0.0215
Unsuccessful Kirschstein-NRSA fellowship applicants
1 -0.6491 0.2337 -1.1071 -0.1911 7.72 0.0055
Other research doctorates
0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.5698 0.1112 -0.7878 -0.3518 26.24 <.0001
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 1.3396 0.3092 0.7336 1.9456 18.77 <.0001
Postdoctoral training 1 0.3198 0.3052 -0.2784 0.9180 1.10 0.2947
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 -0.1213 0.1275 -0.3712 0.1285 0.91 0.3413
9+ 1 -0.3630 0.1910 -0.7374 0.0114 3.61 0.0574
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support
1 -0.0683 0.0612 -0.1882 0.0517 1.24 0.2648
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 -0.1488 0.1392 -0.4217 0.1241 1.14 0.2852
Ranked from 26 to 100 1 -0.1778 0.1287 -0.4299 0.0744 1.91 0.1671
Ranked from 11 to 25 1 -0.1786 0.1834 -0.5380 0.1807 0.95 0.3299
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 6.09 0.0136 LR
Unsuccessful fellowship applicants versus trainees 1 1.97 0.1610 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 10.95 0.0009 LR
Table D.27. Regression Results of the Proportion of Individuals Tenured 8 Years After Receipt of the Doctorate
Model Information
Dataset WORK.EMPLOY_TENURE
Distribution Binomial
Link function Logit
Dependent variable Tenured 8 years after receipt of the doctorate
Observations
Number of observations read 10,659
Number of observations used 2,576
Number of events 1,145
Number of trials 2,576
Missing values 8,083
Response Profile
OrderedValue academic_tenure2 Total
Frequency
1 TENURE 1,145
2 No Tenur 1,431
PROC GENMOD is modeling the probability that Tenured 8 years after receipt of the doctorate = “TENURE.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 2560 2829.2648 1.1052
Scaled deviance 2560 2829.2648 1.1052
Pearson Chi-Square 2560 2583.7395 1.0093
Scaled Pearson X2 2560 2583.7395 1.0093
Log likelihood -1414.6324
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 5.4428 0.6343 4.1996 6.6859 73.64 <.0001
Ph.D. year 1 -0.0682 0.0072 -0.0823 -0.0541 89.39 <.0001
Study group
Kirschstein-NRSA fellows
1 -1.3825 0.2619 -1.8959 -0.8692 27.87 <.0001
Other non-Kirschstein-NRSA postdoctorates
1 -0.5749 0.1394 -0.8481 -0.3017 17.01 <.0001
Kirschstein-NRSA trainees
1 -1.1760 0.2393 -1.6450 -0.7071 24.16 <.0001
Kirschstein-NRSA fellow/trainees
1 -2.1407 0.6159 -3.3478 -0.9337 12.08 0.0005
Unsuccessful Kirschstein-NRSA fellowship applicants
1 -1.1158 0.2385 -1.5832 -0.6484 21.89 <.0001
Other research doctorates
0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.1914 0.0989 -0.3853 0.0025 3.74 0.0530
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 0.5203 0.3258 -0.1182 1.1588 2.55 0.1102
Postdoctoral training 1 -0.5511 0.3345 -1.2068 0.1045 2.71 0.0995
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 0.0974 0.1125 -0.1231 0.3180 0.75 0.3866
9+ 1 0.3156 0.1589 0.0042 0.6270 3.95 0.0470
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support
1 -0.2459 0.1353 -0.5111 0.0193 3.30 0.0692
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 0.0947 0.1229 -0.1462 0.3355 0.59 0.4410
Ranked from 26 to 100 1 0.0331 0.1148 -0.1919 0.2580 0.08 0.7733
Ranked from 11 to 25 1 0.1770 0.1564 -0.1296 0.4836 1.28 0.2579
Top 10 institutions 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 0.47 0.4912 LR
Unsuccessful fellowship applicants versus Kirschstein-NRSA trainees
1 0.04 0.8392 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 7.10 0.0077 LR
Table D.28. Regression Results of the Proportion of Individuals Tenured 12 Years After Receipt of the Doctorate
Model Information
Dataset WORK.EMPLOY_TENURE
Distribution Binomial
Link function Logit
Dependent variable Tenured 12 years after receipt of the doctorate
Observations
Number of observations read 10,659
Number of observations used 2,510
Number of events 1,712
Number of trials 2,510
Missing values 8,149
Response Profile
OrderedValue academic_tenure3 Total
Frequency
1 TENURE 1,712
2 No Tenur 798
PROC GENMOD is modeling the probability that Tenured 12 years after receipt of the doctorate = “TENURE.”
Criteria for Assessing Goodness of Fit
Criterion DF Value Value/DF
Deviance 2494 2567.0399 1.0293
Scaled deviance 2494 2567.0399 1.0293
Pearson Chi-Square 2494 2722.9781 1.0918
Scaled Pearson X2 2494 2722.9781 1.0918
Log likelihood -1283.5199
Analysis of Parameter Estimates
Parameter DF Estimate Standard Error
Wald 95% Confidence
LimitsChi-Square Pr > ChiSq
Intercept 1 12.9184 0.8005 11.3494 14.4874 260.43 <.0001
Ph.D. year 1 -0.1507 0.0093 -0.1689 -0.1325 263.94 <.0001
Study group
Kirschstein-NRSA fellows
1 -0.3455 0.2191 -0.7750 0.0839 2.49 0.1148
Other non-Kirschstein-NRSA postdoctorates
1 -0.0973 0.1580 -0.4070 0.2124 0.38 0.5382
Kirschstein-NRSA trainees
1 -0.5657 0.2045 -0.9666 -0.1648 7.65 0.0057
Kirschstein-NRSA fellow/trainees
1 -0.1231 0.3162 -0.7428 0.4965 0.15 0.6969
Unsuccessful Kirschstein-NRSA fellowship applicants
1 -0.1691 0.2242 -0.6085 0.2703 0.57 0.4508
Other research doctorates
0 0.0000 0.0000 0.0000 0.0000 . .
GenderFemale 1 -0.2440 0.1040 -0.4479 -0.0401 5.50 0.0190
Male 0 0.0000 0.0000 0.0000 0.0000 . .
Expressed postdoctoral plans
Employment 1 0.2358 0.3543 -0.4586 0.9303 0.44 0.5057
Postdoctoral training 1 -0.5067 0.3514 -1.1954 0.1820 2.08 0.1493
Unknown 0 0.0000 0.0000 0.0000 0.0000 . .
Elapsed years enrolled in doctoral program
5–8 1 0.3430 0.1191 0.1095 0.5765 8.29 0.0040
9+ 1 0.1831 0.1702 -0.1505 0.5166 1.16 0.2820
Less than 5 0 0.0000 0.0000 0.0000 0.0000 . .
Received Kirschstein-NRSA predoctoral support
1 -0.0060 0.0633 -0.1301 0.1182 0.01 0.9250
NIH extramural funding (ranking of predoctoral institution)
Ranked 101 or more 1 0.0519 0.1305 -0.2040 0.3077 0.16 0.6912
Ranked from 26 to 100 1 -0.1083 0.1203 -0.3440 0.1274 0.81 0.3677
Ranked from 11 to 25 1 -0.1556 0.1673 -0.4835 0.1723 0.87 0.3523
Top 10 institution 0 0.0000 0.0000 0.0000 0.0000 . .
Scale 0 1.0000 0.0000 1.0000 1.0000
Note: The scale parameter was held fixed.
Contrast Results
Contrast DF Chi-Square Pr > ChiSq Type
Kirschstein-NRSA fellows versus trainees 1 1.13 0.2869 LR
Unsuccessful applicants versus Kirschstein-NRSA trainees
1 3.03 0.0815 LR
Kirschstein-NRSA trainees versus other non-Kirschstein-NRSA postdoctorates
1 6.50 0.0108 LR